content large_stringlengths 0 6.46M | path large_stringlengths 3 331 | license_type large_stringclasses 2 values | repo_name large_stringlengths 5 125 | language large_stringclasses 1 value | is_vendor bool 2 classes | is_generated bool 2 classes | length_bytes int64 4 6.46M | extension large_stringclasses 75 values | text stringlengths 0 6.46M |
|---|---|---|---|---|---|---|---|---|---|
startTime <- Sys.time()
cat(paste0("> Rscript AUC_coexprDist_withFam_sortNoDup_otherTADfile_otherFamFile.R\n"))
options(scipen=100)
buildTable <- TRUE
printAndLog <- function(text, logFile = ""){
cat(text)
cat(text, append =T , file = logFile)
}
suppressPackageStartupMessages(library(foreach, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(doMC, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(ggpubr, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(ggstatsplot, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(ggplot2, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(dplyr, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(flux, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
axisLabSize <- 12
legendSize <- 10
plotTitSize <- 14
mytheme <- theme(
# top, right, bottom and left
plot.margin = unit(c(1.5, 1.5, 1.5, 1.5), "lines"),
plot.title = element_text(hjust = 0.5, face = "bold", size=plotTitSize, vjust=1),
plot.subtitle = element_text(hjust = 0.5, face = "bold", size=plotTitSize-2, vjust=1),
panel.background = element_rect(fill = "white", colour = NA),
panel.border = element_rect(fill = NA, colour = "grey20"),
panel.grid.major = element_line(colour = "grey92"),
panel.grid.minor = element_line(colour = "grey92", size = 0.25),
strip.background = element_rect(fill = "grey85", colour = "grey20"),
#legend.key = element_rect(fill = "white", colour = NA),
axis.line.x = element_line(size = .3, color = "black"),
axis.line.y = element_line(size = .3, color = "black"),
axis.text.y = element_text(color="black", hjust=1,vjust = 0.5, size=axisLabSize),
axis.text.x = element_text(color="black", hjust=0.5,vjust = 1, size=axisLabSize),
axis.title.y = element_text(color="black", size=axisLabSize+1),
axis.title.x = element_text(color="black", size=axisLabSize+1),
legend.text = element_text(size=legendSize),
legend.key.height = unit(1.5,"cm"),
legend.key = element_blank()
)
SSHFS <- F
setDir <- ifelse(SSHFS, "/media/electron", "")
registerDoMC(ifelse(SSHFS, 2, 40))
### HARD CODED
corMethod <- "pearson"
# for plotting:
# look at coexpression ~ distance up to distLimit bp
distLimit <- 500 * 10^3
fitMeth <- "loess"
# nbr of points for loess fit to take the AUC
nbrLoessPoints <- 1000
scatterFontSizeLabel <- 14
scatterFontSizeTitle <- 12
# UPDATE 30.06.2018:
# -> check that always $gene1 < $gene2 before left_join !!!
### RETRIEVE FROM COMMAND LINE
# Rscript AUC_coexprDist_withFam_sortNoDup_otherTADfile.R <TADlist> <dataset>
# Rscript AUC_coexprDist_withFam_sortNoDup_otherTADfile.R <TADlist> <dataset> <family>
# Rscript AUC_coexprDist_withFam_sortNoDup_otherTADfile_otherFamFile.R ENCSR079VIJ_G401_40kb TCGAkich_norm_kich
#### !!! change for otherTADfile: <<< GENE DATA DO NOT CHANGE
# retrieve the sameTAD data frame from:
# file.path("CREATE_SAME_TAD_SORTNODUP", curr_TADlist, "all_TAD_pairs.Rdata")
args <- commandArgs(trailingOnly = TRUE)
stopifnot(length(args) == 2 | length(args) == 3)
if(length(args) == 3) {
txt <- paste0("> Parameters retrieved from command line:\n")
stopifnot(length(args) == 2)
curr_TADlist <- args[1]
curr_dataset <- args[2]
familyData <- args[3]
} else if(length(args) == 2){
curr_TADlist <- args[1]
curr_dataset <- args[2]
txt <- paste0("> Default parameters:\n")
familyData <- "hgnc"
}
outFold <- file.path("AUC_COEXPRDIST_WITHFAM_SORTNODUP", curr_TADlist, paste0(curr_dataset, "_", familyData))
dir.create(outFold, recursive = TRUE)
logFile <- file.path(outFold, paste0("coexpr_dist_withFam_otherTADfile_logFile.txt"))
file.remove(logFile)
printAndLog(txt, logFile)
txt <- paste0("... curr_TADlist = ", curr_TADlist, "\n")
printAndLog(txt, logFile)
txt <- paste0("... curr_dataset = ", curr_dataset, "\n")
printAndLog(txt, logFile)
txt <- paste0("... familyData = ", familyData, "\n")
printAndLog(txt, logFile)
txt <- paste0("> ! Hard-coded parameters:\n")
printAndLog(txt, logFile)
txt <- paste0("... corMethod = ", corMethod, "\n")
printAndLog(txt, logFile)
txt <- paste0("... buildTable = ", as.character(buildTable), "\n")
printAndLog(txt, logFile)
txt <- paste0("... distLimit = ", distLimit, "\n")
printAndLog(txt, logFile)
txt <- paste0("... fitMeth = ", fitMeth, "\n")
printAndLog(txt, logFile)
mycols <- c("same TAD" ="darkorange1" , "diff. TAD"="darkslateblue", "same Fam. + same TAD"="violetred1", "same Fam. + diff. TAD" = "lightskyblue")
sameTADcol <- mycols["same TAD"]
diffTADcol <- mycols["diff. TAD"]
sameFamSameTADcol <- mycols["same Fam. + same TAD"]
sameFamDiffTADcol <- mycols["same Fam. + diff. TAD"]
plotType <- "png"
# myHeight <- ifelse(plotType == "png", 400, 7)
# myWidth <- ifelse(plotType == "png", 600, 10)
myHeight <- ifelse(plotType == "png", 400, 5)
myWidth <- ifelse(plotType == "png", 600, 6)
pipScriptDir <- paste0(setDir, "/mnt/ed4/marie/scripts/TAD_DE_pipeline_v2")
source(paste0(pipScriptDir, "/", "TAD_DE_utils.R"))
utilsDir <- paste0(setDir, "/mnt/ed4/marie/scripts/TAD_DE_pipeline_v2_coreg")
source(file.path(utilsDir, "coreg_utils_ggscatterhist.R"))
### UPDATE 08.01.19 => USE TAD LIST SPECIFIC FILES = TISSUE SPECIFIC FILES
distFile <- file.path("CREATE_DIST_SORTNODUP", curr_TADlist, "all_dist_pairs.Rdata")
stopifnot(file.exists(distFile))
# CHANGED 08.01.19 !!!
coexprFile <- file.path("CREATE_COEXPR_SORTNODUP", curr_TADlist, paste0(curr_dataset), corMethod, "coexprDT.Rdata")
stopifnot(file.exists(coexprFile))
sameTADfile <- file.path("CREATE_SAME_TAD_SORTNODUP", curr_TADlist, "all_TAD_pairs.Rdata")
stopifnot(file.exists(sameTADfile))
# ADDED 08.01.19 to accommodate updated family file
sameFamFolder <- file.path("CREATE_SAME_FAMILY_SORTNODUP", curr_TADlist)
# checking the file comes after (iterating over family and family_short)
stopifnot(dir.exists(sameFamFolder))
sameFamFile <- file.path(sameFamFolder, paste0(familyData, "_family_all_family_pairs.Rdata")) # at least this one should exist !
stopifnot(file.exists(sameFamFile))
dataset_pipDir <- file.path("PIPELINE", "OUTPUT_FOLDER", curr_TADlist, curr_dataset) # used to retrieve gene list
stopifnot(dir.exists(dataset_pipDir))
txt <- paste0("... distFile = ", distFile, "\n")
printAndLog(txt, logFile)
txt <- paste0("... coexprFile = ", coexprFile, "\n")
printAndLog(txt, logFile)
txt <- paste0("... sameTADfile = ", sameTADfile, "\n")
printAndLog(txt, logFile)
txt <- paste0("... sameFamFile = ", sameFamFile, "\n")
printAndLog(txt, logFile)
################################################ DATA PREPARATION
if(buildTable) {
cat(paste0("... load DIST data\t", distFile, "\t", Sys.time(), "\t"))
load(distFile)
cat(paste0(Sys.time(), "\n"))
head(all_dist_pairs)
nrow(all_dist_pairs)
all_dist_pairs$gene1 <- as.character(all_dist_pairs$gene1)
all_dist_pairs$gene2 <- as.character(all_dist_pairs$gene2)
# UPDATE 30.06.2018
stopifnot(all_dist_pairs$gene1 < all_dist_pairs$gene2)
cat(paste0("... load TAD data\t", sameTADfile, "\t", Sys.time(), "\t"))
### =>>> CHANGED HERE FOR OTHER TAD FILE !!!
load(sameTADfile)
cat(paste0(Sys.time(), "\n"))
head(all_TAD_pairs)
nrow(all_TAD_pairs)
all_TAD_pairs$gene1 <- as.character(all_TAD_pairs$gene1)
all_TAD_pairs$gene2 <- as.character(all_TAD_pairs$gene2)
# UPDATE 30.06.2018
stopifnot(all_TAD_pairs$gene1 < all_TAD_pairs$gene2)
cat(paste0("... load COEXPR data\t",coexprFile, "\t", Sys.time(), "\t"))
load(coexprFile)
cat(paste0(Sys.time(), "\n"))
head(coexprDT)
nrow(coexprDT)
coexprDT$gene1 <- as.character(coexprDT$gene1)
coexprDT$gene2 <- as.character(coexprDT$gene2)
all_TAD_pairs$gene2
# UPDATE 30.06.2018
stopifnot(coexprDT$gene1 < coexprDT$gene2)
#============================== RETRIEVE PIPELINE DATA FOR THIS DATASET - USED ONLY FOR GENE LIST
script0_name <- "0_prepGeneData"
geneFile <- file.path(dataset_pipDir, script0_name, "pipeline_geneList.Rdata")
cat(paste0("... load GENELIST file\t",geneFile, "\t", Sys.time(), "\t"))
pipeline_geneList <- eval(parse(text = load(geneFile)))
pipeline_geneList <- as.character(pipeline_geneList)
dataset_dist_pair <- all_dist_pairs[all_dist_pairs$gene1 %in% pipeline_geneList &
all_dist_pairs$gene2 %in% pipeline_geneList,]
dataset_dist_pairs_limit <- dataset_dist_pair[dataset_dist_pair$dist <= distLimit,]
head(dataset_dist_pairs_limit)
nrow(dataset_dist_pairs_limit)
dataset_TAD_pairs <- all_TAD_pairs[all_TAD_pairs$gene1 %in% pipeline_geneList &
all_TAD_pairs$gene2 %in% pipeline_geneList,]
head(dataset_TAD_pairs)
nrow(dataset_TAD_pairs)
# START MERGING DATA
cat(paste0("... merge DIST - TAD data\t", Sys.time(), "\t"))
dataset_dist_TAD_DT <- left_join(dataset_dist_pairs_limit, dataset_TAD_pairs, by=c("gene1", "gene2"))
cat(paste0(Sys.time(), "\n"))
dataset_dist_TAD_DT$sameTAD <- ifelse(is.na(dataset_dist_TAD_DT$region), 0, 1)
}
# all_familyData <- paste0(familyData, c("_family", "_family_short"))
all_familyData <- paste0(familyData, c( "_family_short"))
outFold_save <- outFold
for(i_fam in all_familyData) {
outFold <- file.path(outFold_save, i_fam)
dir.create(outFold, recursive = TRUE)
if(buildTable){
# UPDATE HERE 08.01.19 TO HAVE DATA FROM TISSUE SPECIFIC TAD LIST
sameFamFile <- file.path(sameFamFolder, paste0(i_fam, "_all_family_pairs.Rdata"))
cat(paste0("... load FAMILY data\t", sameFamFile, "\n", Sys.time(), "\t"))
stopifnot(file.exists(sameFamFile))
load(sameFamFile)
cat(paste0(Sys.time(), "\n"))
head(all_family_pairs)
nrow(all_family_pairs)
all_family_pairs$gene1 <- as.character(all_family_pairs$gene1)
all_family_pairs$gene2 <- as.character(all_family_pairs$gene2)
stopifnot(all_family_pairs$gene1 < all_family_pairs$gene2)
dataset_family_pairs <- all_family_pairs[all_family_pairs$gene1 %in% pipeline_geneList &
all_family_pairs$gene2 %in% pipeline_geneList,]
head(dataset_family_pairs)
cat(paste0("... merge FAMILY data\t", Sys.time(), "\t"))
dataset_dist_TAD_fam_DT <- left_join(dataset_dist_TAD_DT, dataset_family_pairs, by=c("gene1", "gene2"))
cat(paste0(Sys.time(), "\n"))
dataset_dist_TAD_fam_DT$sameFamily <- ifelse(is.na(dataset_dist_TAD_fam_DT$family), 0, 1)
cat(paste0("... merge COEXPR data\t", Sys.time(), "\t"))
dataset_dist_TAD_fam_coexpr_DT <- left_join(dataset_dist_TAD_fam_DT, coexprDT, by=c("gene1", "gene2"))
cat(paste0(Sys.time(), "\n"))
allData_dt <- dataset_dist_TAD_fam_coexpr_DT
allData_dt$region <- NULL
allData_dt$family <- NULL
allData_dt <- na.omit(allData_dt)
# outFile <-file.path(outFold, paste0(i_fam, "_", "allData_dt.Rdata"))
outFile <-file.path(outFold, paste0( "allData_dt.Rdata"))
save(allData_dt, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
} else{
# outFile <-file.path(outFold, paste0(i_fam, "_", "allData_dt.Rdata"))
outFile <-file.path(outFold, paste0( "allData_dt.Rdata"))
load(outFile)
# load("/media/electron/mnt/ed4/marie/scripts/TAD_DE_pipeline_v2_TopDom/COEXPR_DIST_v3/TCGAcrc_msi_mss_hgnc/hgnc_family_allData_dt.Rdata")
}
nrow(allData_dt)
allData_dt$dist_kb <- allData_dt$dist/1000
allData_dt$curve1 <- ifelse(allData_dt$sameTAD == "0", "diff. TAD", "same TAD")
allData_dt$curve2 <- ifelse(allData_dt$sameFamily == "0", NA,
ifelse(allData_dt$sameTAD == "0", "same Fam. + diff. TAD", "same Fam. + same TAD"))
sameTAD_DT <- allData_dt[allData_dt$sameTAD == 1,c("gene1", "gene2", "coexpr", "dist", "dist_kb")]
sameTAD_DT <- na.omit(sameTAD_DT)
sameTAD_DT <- sameTAD_DT[order(sameTAD_DT$dist_kb),]
sameTAD_DT$nPair <- 1:nrow(sameTAD_DT)
sameTAD_DT$label <- "same TAD"
diffTAD_DT <- allData_dt[allData_dt$sameTAD == 0,c("gene1", "gene2", "coexpr", "dist", "dist_kb")]
diffTAD_DT <- na.omit(diffTAD_DT)
diffTAD_DT <- diffTAD_DT[order(diffTAD_DT$dist_kb),]
diffTAD_DT$nPair <- 1:nrow(diffTAD_DT)
diffTAD_DT$label <- "diff. TAD"
sameFam_sameTAD_DT <- allData_dt[allData_dt$sameFamily == 1 & allData_dt$sameTAD == 1 ,c("gene1", "gene2", "coexpr", "dist", "dist_kb")]
sameFam_sameTAD_DT <- na.omit(sameFam_sameTAD_DT)
sameFam_sameTAD_DT <- sameFam_sameTAD_DT[order(sameFam_sameTAD_DT$dist_kb),]
sameFam_sameTAD_DT$nPair <- 1:nrow(sameFam_sameTAD_DT)
sameFam_sameTAD_DT$label <- "same Fam. + same TAD"
sameFam_diffTAD_DT <- allData_dt[allData_dt$sameFamily == 1 & allData_dt$sameTAD == 0 ,c("gene1", "gene2", "coexpr", "dist", "dist_kb")]
sameFam_diffTAD_DT <- na.omit(sameFam_diffTAD_DT)
sameFam_diffTAD_DT <- sameFam_diffTAD_DT[order(sameFam_diffTAD_DT$dist_kb),]
sameFam_diffTAD_DT$nPair <- 1:nrow(sameFam_diffTAD_DT)
sameFam_diffTAD_DT$label <- "same Fam. + diff. TAD"
stopifnot(is.numeric(sameTAD_DT$dist[1]))
stopifnot(is.numeric(sameTAD_DT$coexpr[1]))
stopifnot(is.numeric(diffTAD_DT$dist[1]))
stopifnot(is.numeric(diffTAD_DT$coexpr[1]))
stopifnot(is.numeric(sameFam_sameTAD_DT$dist[1]))
stopifnot(is.numeric(sameFam_sameTAD_DT$coexpr[1]))
stopifnot(is.numeric(sameFam_diffTAD_DT$dist[1]))
stopifnot(is.numeric(sameFam_diffTAD_DT$coexpr[1]))
#***
if(fitMeth == "loess") {
my_ylab <- paste0("Gene pair coexpression (", corMethod, ", qqnormDT)")
my_xlab <- paste0("Distance between the 2 genes (kb)")
my_sub <- paste0(curr_dataset)
# PREDICT WITH ORIGINAL DISTANCE VALUES
my_xlab <- paste0("Distance between the 2 genes (bp)")
diffTAD_mod <- loess(coexpr ~ dist, data = diffTAD_DT)
sameTAD_mod <- loess(coexpr ~ dist, data = sameTAD_DT)
smooth_vals_sameTAD <- predict(sameTAD_mod, sort(sameTAD_DT$dist))
smooth_vals_diffTAD <- predict(diffTAD_mod, sort(diffTAD_DT$dist))
auc_diffTAD_obsDist <- auc(x = sort(diffTAD_DT$dist), y = smooth_vals_diffTAD)
auc_sameTAD_obsDist <- auc(x = sort(sameTAD_DT$dist), y = smooth_vals_sameTAD)
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameTAD_diffTAD_loessFit_originalDist", ".", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(allData_dt$dist),
ylim = range(c(smooth_vals_sameTAD, smooth_vals_diffTAD)),
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = "observed distance values", side = 3)
lines( x = sort(sameTAD_DT$dist), y = smooth_vals_sameTAD, col = sameTADcol)
lines( x = sort(diffTAD_DT$dist), y = smooth_vals_diffTAD, col = diffTADcol)
legend("topright",
legend=c(paste0("sameTAD\n(AUC=", round(auc_sameTAD_obsDist, 2), ")"), paste0("diffTAD\n(AUC=", round(auc_diffTAD_obsDist, 2))),
col = c(sameTADcol, diffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
# PREDICT WITH DISTANCE VECTOR
distVect <- seq(from=0, to = distLimit, length.out = nbrLoessPoints)
smooth_vals_sameTAD_distVect <- predict(sameTAD_mod, distVect)
smooth_vals_diffTAD_distVect <- predict(diffTAD_mod, distVect)
auc_diffTAD_distVect <- auc(x = distVect, y = smooth_vals_diffTAD_distVect)
auc_sameTAD_distVect <- auc(x = distVect, y = smooth_vals_sameTAD_distVect)
outFile <- file.path(outFold, "auc_diffTAD_distVect.Rdata")
save(auc_diffTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "auc_sameTAD_distVect.Rdata")
save(auc_sameTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "smooth_vals_sameTAD_distVect.Rdata")
save(smooth_vals_sameTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "smooth_vals_diffTAD_distVect.Rdata")
save(smooth_vals_diffTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "distVect.Rdata")
save(distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameTAD_diffTAD_loessFit_vectDist.", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(distVect),
ylim = range(c(na.omit(smooth_vals_sameTAD_distVect), na.omit(smooth_vals_diffTAD_distVect))),
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = paste0("distance values seq from 0 to ", distLimit, " (# points = ", nbrLoessPoints, ")"), side = 3)
lines( x = distVect, y = smooth_vals_sameTAD_distVect, col = sameTADcol)
lines( x = distVect, y = smooth_vals_diffTAD_distVect, col = diffTADcol)
legend("topright",
legend=c(paste0("sameTAD\n(AUC=", round(auc_sameTAD_distVect, 2), ")"), paste0("diffTAD\n(AUC=", round(auc_diffTAD_distVect, 2))),
col = c(sameTADcol, diffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
################################ DO THE SAME FOR SAME FAM SAME TAD VS. SAME FAM DIFF TAD
sameFamDiffTAD_mod <- loess(coexpr ~ dist, data = sameFam_diffTAD_DT)
sameFamSameTAD_mod <- loess(coexpr ~ dist, data = sameFam_sameTAD_DT)
smooth_vals_sameFamSameTAD <- predict(sameFamSameTAD_mod, sort(sameFam_sameTAD_DT$dist))
smooth_vals_sameFamDiffTAD <- predict(sameFamDiffTAD_mod, sort(sameFam_diffTAD_DT$dist))
auc_sameFamDiffTAD_obsDist <- auc(x = sort(sameFam_diffTAD_DT$dist), y = smooth_vals_sameFamDiffTAD)
auc_sameFamSameTAD_obsDist <- auc(x = sort(sameFam_sameTAD_DT$dist), y = smooth_vals_sameFamSameTAD)
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameFamSameTAD_sameFamDiffTAD_loessFit_originalDist", ".", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(allData_dt$dist),
ylim = range(c(smooth_vals_sameFamSameTAD, smooth_vals_sameFamDiffTAD)),
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = "observed distance values", side = 3)
lines( x = sort(sameFam_sameTAD_DT$dist), y = smooth_vals_sameFamSameTAD, col = sameFamSameTADcol)
lines( x = sort(sameFam_diffTAD_DT$dist), y = smooth_vals_sameFamDiffTAD, col = sameFamDiffTADcol)
legend("topright",
legend=c(paste0("sameFamSameTAD\n(AUC=", round(auc_sameFamSameTAD_obsDist, 2), ")"),
paste0("sameFamDiffTAD\n(AUC=", round(auc_sameFamDiffTAD_obsDist, 2))),
col = c(sameFamSameTADcol, sameFamDiffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
# PREDICT WITH DISTANCE VECTOR
smooth_vals_sameFamSameTAD_distVect <- predict(sameFamSameTAD_mod, distVect)
smooth_vals_sameFamDiffTAD_distVect <- predict(sameFamDiffTAD_mod, distVect)
auc_sameFamDiffTAD_distVect <- auc(x = distVect, y = smooth_vals_sameFamDiffTAD_distVect)
auc_sameFamSameTAD_distVect <- auc(x = distVect, y = smooth_vals_sameFamSameTAD_distVect)
outFile <- file.path(outFold, "diffTAD_mod.Rdata")
save(diffTAD_mod, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "sameFamDiffTAD_mod.Rdata")
save(sameFamDiffTAD_mod, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "sameTAD_mod.Rdata")
save(sameTAD_mod, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "sameFamSameTAD_mod.Rdata")
save(sameFamSameTAD_mod, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
sameFamDiffTAD_obsDist <- sameFam_diffTAD_DT$dist
outFile <- file.path(outFold, "sameFamDiffTAD_obsDist.Rdata")
save(sameFamDiffTAD_obsDist, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
sameFamSameTAD_obsDist <- sameFam_sameTAD_DT$dist
outFile <- file.path(outFold, "sameFamSameTAD_obsDist.Rdata")
save(sameFamSameTAD_obsDist, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "auc_sameFamDiffTAD_distVect.Rdata")
save(auc_sameFamDiffTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "auc_sameFamSameTAD_distVect.Rdata")
save(auc_sameFamSameTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "smooth_vals_sameFamSameTAD_distVect.Rdata")
save(smooth_vals_sameFamSameTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "smooth_vals_sameFamDiffTAD_distVect.Rdata")
save(smooth_vals_sameFamDiffTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameFamSameTAD_sameFamDiffTAD_loessFit_vectDist.", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(distVect),
ylim = range(c(na.omit(smooth_vals_sameFamSameTAD_distVect), na.omit(smooth_vals_sameFamDiffTAD_distVect))),
# xlab="",
# ylab="",
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = paste0("distance values seq from 0 to ", distLimit, " (# points = ", nbrLoessPoints, ")"), side = 3)
lines( x = distVect, y = smooth_vals_sameFamSameTAD_distVect, col = sameFamSameTADcol)
lines( x = distVect, y = smooth_vals_sameFamDiffTAD_distVect, col = sameFamDiffTADcol)
legend("topright",
legend=c(paste0("sameFamSameTAD\n(AUC=", round(auc_sameFamSameTAD_distVect, 2), ")"),
paste0("sameFamDiffTAD\n(AUC=", round(auc_sameFamDiffTAD_distVect, 2))),
col = c(sameFamSameTADcol, sameFamDiffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
################################################################
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameTAD_diffTAD_sameFamSameTAD_sameFamDiffTAD_loessFit_vectDist.", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(distVect),
ylim = range(c(na.omit(smooth_vals_sameTAD_distVect),
na.omit(smooth_vals_sameFamSameTAD_distVect),
na.omit(smooth_vals_diffTAD_distVect),
na.omit(smooth_vals_sameFamDiffTAD_distVect))),
# xlab="",
# ylab="",
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = paste0("distance values seq from 0 to ", distLimit, " (# points = ", nbrLoessPoints, ")"), side = 3)
lines( x = distVect, y = smooth_vals_sameFamSameTAD_distVect, col = sameFamSameTADcol)
lines( x = distVect, y = smooth_vals_sameFamDiffTAD_distVect, col = sameFamDiffTADcol)
lines( x = distVect, y = smooth_vals_sameTAD_distVect, col = sameTADcol)
lines( x = distVect, y = smooth_vals_diffTAD_distVect, col = diffTADcol)
legend("topright",
legend=c(paste0("sameTAD\n(AUC=", round(auc_sameTAD_distVect, 2), ")"),
paste0("diffTAD\n(AUC=", round(auc_diffTAD_distVect, 2)),
paste0("sameFamSameTAD\n(AUC=", round(auc_sameFamSameTAD_distVect, 2), ")"),
paste0("sameFamDiffTAD\n(AUC=", round(auc_sameFamDiffTAD_distVect, 2))),
col = c(sameTADcol, diffTADcol,
sameFamSameTADcol, sameFamDiffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
################################################################
auc_values <- list(
auc_diffTAD_distVect = auc_diffTAD_distVect,
auc_sameTAD_distVect = auc_sameTAD_distVect,
auc_ratio_same_over_diff_distVect = auc_sameTAD_distVect/auc_diffTAD_distVect,
auc_diffTAD_obsDist = auc_diffTAD_obsDist,
auc_sameTAD_obsDist = auc_sameTAD_obsDist,
auc_ratio_same_over_diff_obsDist = auc_sameTAD_distVect/auc_diffTAD_obsDist,
auc_sameFamDiffTAD_distVect = auc_sameFamDiffTAD_distVect,
auc_sameFamSameTAD_distVect = auc_sameFamSameTAD_distVect,
auc_ratio_sameFam_same_over_diff_distVect = auc_sameFamSameTAD_distVect/auc_sameFamDiffTAD_distVect,
auc_sameFamDiffTAD_obsDist = auc_sameFamDiffTAD_obsDist,
auc_sameFamSameTAD_obsDist = auc_sameFamSameTAD_obsDist,
auc_ratio_sameFam_same_over_diff_obsDist = auc_sameFamSameTAD_distVect/auc_sameFamDiffTAD_obsDist
)
outFile <- file.path(outFold, paste0("auc_values.Rdata"))
save(auc_values, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
} else{
stop("only loess implemented yet\n")
}
} # end iterating over family data
######################################################################################
######################################################################################
######################################################################################
cat(paste0("... written: ", logFile, "\n"))
######################################################################################
cat("*** DONE\n")
cat(paste0(startTime, "\n", Sys.time(), "\n"))
| /AUC_coexprDist_withFam_sortNoDup_otherTADfile_otherFamFile.R | no_license | marzuf/CORRECT_Yuanlong_Cancer_HiC_data_TAD_DA | R | false | false | 26,452 | r | startTime <- Sys.time()
cat(paste0("> Rscript AUC_coexprDist_withFam_sortNoDup_otherTADfile_otherFamFile.R\n"))
options(scipen=100)
buildTable <- TRUE
printAndLog <- function(text, logFile = ""){
cat(text)
cat(text, append =T , file = logFile)
}
suppressPackageStartupMessages(library(foreach, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(doMC, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(ggpubr, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(ggstatsplot, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(ggplot2, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(dplyr, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
suppressPackageStartupMessages(library(flux, warn.conflicts = FALSE, quietly = TRUE, verbose = FALSE))
axisLabSize <- 12
legendSize <- 10
plotTitSize <- 14
mytheme <- theme(
# top, right, bottom and left
plot.margin = unit(c(1.5, 1.5, 1.5, 1.5), "lines"),
plot.title = element_text(hjust = 0.5, face = "bold", size=plotTitSize, vjust=1),
plot.subtitle = element_text(hjust = 0.5, face = "bold", size=plotTitSize-2, vjust=1),
panel.background = element_rect(fill = "white", colour = NA),
panel.border = element_rect(fill = NA, colour = "grey20"),
panel.grid.major = element_line(colour = "grey92"),
panel.grid.minor = element_line(colour = "grey92", size = 0.25),
strip.background = element_rect(fill = "grey85", colour = "grey20"),
#legend.key = element_rect(fill = "white", colour = NA),
axis.line.x = element_line(size = .3, color = "black"),
axis.line.y = element_line(size = .3, color = "black"),
axis.text.y = element_text(color="black", hjust=1,vjust = 0.5, size=axisLabSize),
axis.text.x = element_text(color="black", hjust=0.5,vjust = 1, size=axisLabSize),
axis.title.y = element_text(color="black", size=axisLabSize+1),
axis.title.x = element_text(color="black", size=axisLabSize+1),
legend.text = element_text(size=legendSize),
legend.key.height = unit(1.5,"cm"),
legend.key = element_blank()
)
SSHFS <- F
setDir <- ifelse(SSHFS, "/media/electron", "")
registerDoMC(ifelse(SSHFS, 2, 40))
### HARD CODED
corMethod <- "pearson"
# for plotting:
# look at coexpression ~ distance up to distLimit bp
distLimit <- 500 * 10^3
fitMeth <- "loess"
# nbr of points for loess fit to take the AUC
nbrLoessPoints <- 1000
scatterFontSizeLabel <- 14
scatterFontSizeTitle <- 12
# UPDATE 30.06.2018:
# -> check that always $gene1 < $gene2 before left_join !!!
### RETRIEVE FROM COMMAND LINE
# Rscript AUC_coexprDist_withFam_sortNoDup_otherTADfile.R <TADlist> <dataset>
# Rscript AUC_coexprDist_withFam_sortNoDup_otherTADfile.R <TADlist> <dataset> <family>
# Rscript AUC_coexprDist_withFam_sortNoDup_otherTADfile_otherFamFile.R ENCSR079VIJ_G401_40kb TCGAkich_norm_kich
#### !!! change for otherTADfile: <<< GENE DATA DO NOT CHANGE
# retrieve the sameTAD data frame from:
# file.path("CREATE_SAME_TAD_SORTNODUP", curr_TADlist, "all_TAD_pairs.Rdata")
args <- commandArgs(trailingOnly = TRUE)
stopifnot(length(args) == 2 | length(args) == 3)
if(length(args) == 3) {
txt <- paste0("> Parameters retrieved from command line:\n")
stopifnot(length(args) == 2)
curr_TADlist <- args[1]
curr_dataset <- args[2]
familyData <- args[3]
} else if(length(args) == 2){
curr_TADlist <- args[1]
curr_dataset <- args[2]
txt <- paste0("> Default parameters:\n")
familyData <- "hgnc"
}
outFold <- file.path("AUC_COEXPRDIST_WITHFAM_SORTNODUP", curr_TADlist, paste0(curr_dataset, "_", familyData))
dir.create(outFold, recursive = TRUE)
logFile <- file.path(outFold, paste0("coexpr_dist_withFam_otherTADfile_logFile.txt"))
file.remove(logFile)
printAndLog(txt, logFile)
txt <- paste0("... curr_TADlist = ", curr_TADlist, "\n")
printAndLog(txt, logFile)
txt <- paste0("... curr_dataset = ", curr_dataset, "\n")
printAndLog(txt, logFile)
txt <- paste0("... familyData = ", familyData, "\n")
printAndLog(txt, logFile)
txt <- paste0("> ! Hard-coded parameters:\n")
printAndLog(txt, logFile)
txt <- paste0("... corMethod = ", corMethod, "\n")
printAndLog(txt, logFile)
txt <- paste0("... buildTable = ", as.character(buildTable), "\n")
printAndLog(txt, logFile)
txt <- paste0("... distLimit = ", distLimit, "\n")
printAndLog(txt, logFile)
txt <- paste0("... fitMeth = ", fitMeth, "\n")
printAndLog(txt, logFile)
mycols <- c("same TAD" ="darkorange1" , "diff. TAD"="darkslateblue", "same Fam. + same TAD"="violetred1", "same Fam. + diff. TAD" = "lightskyblue")
sameTADcol <- mycols["same TAD"]
diffTADcol <- mycols["diff. TAD"]
sameFamSameTADcol <- mycols["same Fam. + same TAD"]
sameFamDiffTADcol <- mycols["same Fam. + diff. TAD"]
plotType <- "png"
# myHeight <- ifelse(plotType == "png", 400, 7)
# myWidth <- ifelse(plotType == "png", 600, 10)
myHeight <- ifelse(plotType == "png", 400, 5)
myWidth <- ifelse(plotType == "png", 600, 6)
pipScriptDir <- paste0(setDir, "/mnt/ed4/marie/scripts/TAD_DE_pipeline_v2")
source(paste0(pipScriptDir, "/", "TAD_DE_utils.R"))
utilsDir <- paste0(setDir, "/mnt/ed4/marie/scripts/TAD_DE_pipeline_v2_coreg")
source(file.path(utilsDir, "coreg_utils_ggscatterhist.R"))
### UPDATE 08.01.19 => USE TAD LIST SPECIFIC FILES = TISSUE SPECIFIC FILES
distFile <- file.path("CREATE_DIST_SORTNODUP", curr_TADlist, "all_dist_pairs.Rdata")
stopifnot(file.exists(distFile))
# CHANGED 08.01.19 !!!
coexprFile <- file.path("CREATE_COEXPR_SORTNODUP", curr_TADlist, paste0(curr_dataset), corMethod, "coexprDT.Rdata")
stopifnot(file.exists(coexprFile))
sameTADfile <- file.path("CREATE_SAME_TAD_SORTNODUP", curr_TADlist, "all_TAD_pairs.Rdata")
stopifnot(file.exists(sameTADfile))
# ADDED 08.01.19 to accommodate updated family file
sameFamFolder <- file.path("CREATE_SAME_FAMILY_SORTNODUP", curr_TADlist)
# checking the file comes after (iterating over family and family_short)
stopifnot(dir.exists(sameFamFolder))
sameFamFile <- file.path(sameFamFolder, paste0(familyData, "_family_all_family_pairs.Rdata")) # at least this one should exist !
stopifnot(file.exists(sameFamFile))
dataset_pipDir <- file.path("PIPELINE", "OUTPUT_FOLDER", curr_TADlist, curr_dataset) # used to retrieve gene list
stopifnot(dir.exists(dataset_pipDir))
txt <- paste0("... distFile = ", distFile, "\n")
printAndLog(txt, logFile)
txt <- paste0("... coexprFile = ", coexprFile, "\n")
printAndLog(txt, logFile)
txt <- paste0("... sameTADfile = ", sameTADfile, "\n")
printAndLog(txt, logFile)
txt <- paste0("... sameFamFile = ", sameFamFile, "\n")
printAndLog(txt, logFile)
################################################ DATA PREPARATION
if(buildTable) {
cat(paste0("... load DIST data\t", distFile, "\t", Sys.time(), "\t"))
load(distFile)
cat(paste0(Sys.time(), "\n"))
head(all_dist_pairs)
nrow(all_dist_pairs)
all_dist_pairs$gene1 <- as.character(all_dist_pairs$gene1)
all_dist_pairs$gene2 <- as.character(all_dist_pairs$gene2)
# UPDATE 30.06.2018
stopifnot(all_dist_pairs$gene1 < all_dist_pairs$gene2)
cat(paste0("... load TAD data\t", sameTADfile, "\t", Sys.time(), "\t"))
### =>>> CHANGED HERE FOR OTHER TAD FILE !!!
load(sameTADfile)
cat(paste0(Sys.time(), "\n"))
head(all_TAD_pairs)
nrow(all_TAD_pairs)
all_TAD_pairs$gene1 <- as.character(all_TAD_pairs$gene1)
all_TAD_pairs$gene2 <- as.character(all_TAD_pairs$gene2)
# UPDATE 30.06.2018
stopifnot(all_TAD_pairs$gene1 < all_TAD_pairs$gene2)
cat(paste0("... load COEXPR data\t",coexprFile, "\t", Sys.time(), "\t"))
load(coexprFile)
cat(paste0(Sys.time(), "\n"))
head(coexprDT)
nrow(coexprDT)
coexprDT$gene1 <- as.character(coexprDT$gene1)
coexprDT$gene2 <- as.character(coexprDT$gene2)
all_TAD_pairs$gene2
# UPDATE 30.06.2018
stopifnot(coexprDT$gene1 < coexprDT$gene2)
#============================== RETRIEVE PIPELINE DATA FOR THIS DATASET - USED ONLY FOR GENE LIST
script0_name <- "0_prepGeneData"
geneFile <- file.path(dataset_pipDir, script0_name, "pipeline_geneList.Rdata")
cat(paste0("... load GENELIST file\t",geneFile, "\t", Sys.time(), "\t"))
pipeline_geneList <- eval(parse(text = load(geneFile)))
pipeline_geneList <- as.character(pipeline_geneList)
dataset_dist_pair <- all_dist_pairs[all_dist_pairs$gene1 %in% pipeline_geneList &
all_dist_pairs$gene2 %in% pipeline_geneList,]
dataset_dist_pairs_limit <- dataset_dist_pair[dataset_dist_pair$dist <= distLimit,]
head(dataset_dist_pairs_limit)
nrow(dataset_dist_pairs_limit)
dataset_TAD_pairs <- all_TAD_pairs[all_TAD_pairs$gene1 %in% pipeline_geneList &
all_TAD_pairs$gene2 %in% pipeline_geneList,]
head(dataset_TAD_pairs)
nrow(dataset_TAD_pairs)
# START MERGING DATA
cat(paste0("... merge DIST - TAD data\t", Sys.time(), "\t"))
dataset_dist_TAD_DT <- left_join(dataset_dist_pairs_limit, dataset_TAD_pairs, by=c("gene1", "gene2"))
cat(paste0(Sys.time(), "\n"))
dataset_dist_TAD_DT$sameTAD <- ifelse(is.na(dataset_dist_TAD_DT$region), 0, 1)
}
# all_familyData <- paste0(familyData, c("_family", "_family_short"))
all_familyData <- paste0(familyData, c( "_family_short"))
outFold_save <- outFold
for(i_fam in all_familyData) {
outFold <- file.path(outFold_save, i_fam)
dir.create(outFold, recursive = TRUE)
if(buildTable){
# UPDATE HERE 08.01.19 TO HAVE DATA FROM TISSUE SPECIFIC TAD LIST
sameFamFile <- file.path(sameFamFolder, paste0(i_fam, "_all_family_pairs.Rdata"))
cat(paste0("... load FAMILY data\t", sameFamFile, "\n", Sys.time(), "\t"))
stopifnot(file.exists(sameFamFile))
load(sameFamFile)
cat(paste0(Sys.time(), "\n"))
head(all_family_pairs)
nrow(all_family_pairs)
all_family_pairs$gene1 <- as.character(all_family_pairs$gene1)
all_family_pairs$gene2 <- as.character(all_family_pairs$gene2)
stopifnot(all_family_pairs$gene1 < all_family_pairs$gene2)
dataset_family_pairs <- all_family_pairs[all_family_pairs$gene1 %in% pipeline_geneList &
all_family_pairs$gene2 %in% pipeline_geneList,]
head(dataset_family_pairs)
cat(paste0("... merge FAMILY data\t", Sys.time(), "\t"))
dataset_dist_TAD_fam_DT <- left_join(dataset_dist_TAD_DT, dataset_family_pairs, by=c("gene1", "gene2"))
cat(paste0(Sys.time(), "\n"))
dataset_dist_TAD_fam_DT$sameFamily <- ifelse(is.na(dataset_dist_TAD_fam_DT$family), 0, 1)
cat(paste0("... merge COEXPR data\t", Sys.time(), "\t"))
dataset_dist_TAD_fam_coexpr_DT <- left_join(dataset_dist_TAD_fam_DT, coexprDT, by=c("gene1", "gene2"))
cat(paste0(Sys.time(), "\n"))
allData_dt <- dataset_dist_TAD_fam_coexpr_DT
allData_dt$region <- NULL
allData_dt$family <- NULL
allData_dt <- na.omit(allData_dt)
# outFile <-file.path(outFold, paste0(i_fam, "_", "allData_dt.Rdata"))
outFile <-file.path(outFold, paste0( "allData_dt.Rdata"))
save(allData_dt, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
} else{
# outFile <-file.path(outFold, paste0(i_fam, "_", "allData_dt.Rdata"))
outFile <-file.path(outFold, paste0( "allData_dt.Rdata"))
load(outFile)
# load("/media/electron/mnt/ed4/marie/scripts/TAD_DE_pipeline_v2_TopDom/COEXPR_DIST_v3/TCGAcrc_msi_mss_hgnc/hgnc_family_allData_dt.Rdata")
}
nrow(allData_dt)
allData_dt$dist_kb <- allData_dt$dist/1000
allData_dt$curve1 <- ifelse(allData_dt$sameTAD == "0", "diff. TAD", "same TAD")
allData_dt$curve2 <- ifelse(allData_dt$sameFamily == "0", NA,
ifelse(allData_dt$sameTAD == "0", "same Fam. + diff. TAD", "same Fam. + same TAD"))
sameTAD_DT <- allData_dt[allData_dt$sameTAD == 1,c("gene1", "gene2", "coexpr", "dist", "dist_kb")]
sameTAD_DT <- na.omit(sameTAD_DT)
sameTAD_DT <- sameTAD_DT[order(sameTAD_DT$dist_kb),]
sameTAD_DT$nPair <- 1:nrow(sameTAD_DT)
sameTAD_DT$label <- "same TAD"
diffTAD_DT <- allData_dt[allData_dt$sameTAD == 0,c("gene1", "gene2", "coexpr", "dist", "dist_kb")]
diffTAD_DT <- na.omit(diffTAD_DT)
diffTAD_DT <- diffTAD_DT[order(diffTAD_DT$dist_kb),]
diffTAD_DT$nPair <- 1:nrow(diffTAD_DT)
diffTAD_DT$label <- "diff. TAD"
sameFam_sameTAD_DT <- allData_dt[allData_dt$sameFamily == 1 & allData_dt$sameTAD == 1 ,c("gene1", "gene2", "coexpr", "dist", "dist_kb")]
sameFam_sameTAD_DT <- na.omit(sameFam_sameTAD_DT)
sameFam_sameTAD_DT <- sameFam_sameTAD_DT[order(sameFam_sameTAD_DT$dist_kb),]
sameFam_sameTAD_DT$nPair <- 1:nrow(sameFam_sameTAD_DT)
sameFam_sameTAD_DT$label <- "same Fam. + same TAD"
sameFam_diffTAD_DT <- allData_dt[allData_dt$sameFamily == 1 & allData_dt$sameTAD == 0 ,c("gene1", "gene2", "coexpr", "dist", "dist_kb")]
sameFam_diffTAD_DT <- na.omit(sameFam_diffTAD_DT)
sameFam_diffTAD_DT <- sameFam_diffTAD_DT[order(sameFam_diffTAD_DT$dist_kb),]
sameFam_diffTAD_DT$nPair <- 1:nrow(sameFam_diffTAD_DT)
sameFam_diffTAD_DT$label <- "same Fam. + diff. TAD"
stopifnot(is.numeric(sameTAD_DT$dist[1]))
stopifnot(is.numeric(sameTAD_DT$coexpr[1]))
stopifnot(is.numeric(diffTAD_DT$dist[1]))
stopifnot(is.numeric(diffTAD_DT$coexpr[1]))
stopifnot(is.numeric(sameFam_sameTAD_DT$dist[1]))
stopifnot(is.numeric(sameFam_sameTAD_DT$coexpr[1]))
stopifnot(is.numeric(sameFam_diffTAD_DT$dist[1]))
stopifnot(is.numeric(sameFam_diffTAD_DT$coexpr[1]))
#***
if(fitMeth == "loess") {
my_ylab <- paste0("Gene pair coexpression (", corMethod, ", qqnormDT)")
my_xlab <- paste0("Distance between the 2 genes (kb)")
my_sub <- paste0(curr_dataset)
# PREDICT WITH ORIGINAL DISTANCE VALUES
my_xlab <- paste0("Distance between the 2 genes (bp)")
diffTAD_mod <- loess(coexpr ~ dist, data = diffTAD_DT)
sameTAD_mod <- loess(coexpr ~ dist, data = sameTAD_DT)
smooth_vals_sameTAD <- predict(sameTAD_mod, sort(sameTAD_DT$dist))
smooth_vals_diffTAD <- predict(diffTAD_mod, sort(diffTAD_DT$dist))
auc_diffTAD_obsDist <- auc(x = sort(diffTAD_DT$dist), y = smooth_vals_diffTAD)
auc_sameTAD_obsDist <- auc(x = sort(sameTAD_DT$dist), y = smooth_vals_sameTAD)
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameTAD_diffTAD_loessFit_originalDist", ".", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(allData_dt$dist),
ylim = range(c(smooth_vals_sameTAD, smooth_vals_diffTAD)),
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = "observed distance values", side = 3)
lines( x = sort(sameTAD_DT$dist), y = smooth_vals_sameTAD, col = sameTADcol)
lines( x = sort(diffTAD_DT$dist), y = smooth_vals_diffTAD, col = diffTADcol)
legend("topright",
legend=c(paste0("sameTAD\n(AUC=", round(auc_sameTAD_obsDist, 2), ")"), paste0("diffTAD\n(AUC=", round(auc_diffTAD_obsDist, 2))),
col = c(sameTADcol, diffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
# PREDICT WITH DISTANCE VECTOR
distVect <- seq(from=0, to = distLimit, length.out = nbrLoessPoints)
smooth_vals_sameTAD_distVect <- predict(sameTAD_mod, distVect)
smooth_vals_diffTAD_distVect <- predict(diffTAD_mod, distVect)
auc_diffTAD_distVect <- auc(x = distVect, y = smooth_vals_diffTAD_distVect)
auc_sameTAD_distVect <- auc(x = distVect, y = smooth_vals_sameTAD_distVect)
outFile <- file.path(outFold, "auc_diffTAD_distVect.Rdata")
save(auc_diffTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "auc_sameTAD_distVect.Rdata")
save(auc_sameTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "smooth_vals_sameTAD_distVect.Rdata")
save(smooth_vals_sameTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "smooth_vals_diffTAD_distVect.Rdata")
save(smooth_vals_diffTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "distVect.Rdata")
save(distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameTAD_diffTAD_loessFit_vectDist.", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(distVect),
ylim = range(c(na.omit(smooth_vals_sameTAD_distVect), na.omit(smooth_vals_diffTAD_distVect))),
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = paste0("distance values seq from 0 to ", distLimit, " (# points = ", nbrLoessPoints, ")"), side = 3)
lines( x = distVect, y = smooth_vals_sameTAD_distVect, col = sameTADcol)
lines( x = distVect, y = smooth_vals_diffTAD_distVect, col = diffTADcol)
legend("topright",
legend=c(paste0("sameTAD\n(AUC=", round(auc_sameTAD_distVect, 2), ")"), paste0("diffTAD\n(AUC=", round(auc_diffTAD_distVect, 2))),
col = c(sameTADcol, diffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
################################ DO THE SAME FOR SAME FAM SAME TAD VS. SAME FAM DIFF TAD
sameFamDiffTAD_mod <- loess(coexpr ~ dist, data = sameFam_diffTAD_DT)
sameFamSameTAD_mod <- loess(coexpr ~ dist, data = sameFam_sameTAD_DT)
smooth_vals_sameFamSameTAD <- predict(sameFamSameTAD_mod, sort(sameFam_sameTAD_DT$dist))
smooth_vals_sameFamDiffTAD <- predict(sameFamDiffTAD_mod, sort(sameFam_diffTAD_DT$dist))
auc_sameFamDiffTAD_obsDist <- auc(x = sort(sameFam_diffTAD_DT$dist), y = smooth_vals_sameFamDiffTAD)
auc_sameFamSameTAD_obsDist <- auc(x = sort(sameFam_sameTAD_DT$dist), y = smooth_vals_sameFamSameTAD)
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameFamSameTAD_sameFamDiffTAD_loessFit_originalDist", ".", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(allData_dt$dist),
ylim = range(c(smooth_vals_sameFamSameTAD, smooth_vals_sameFamDiffTAD)),
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = "observed distance values", side = 3)
lines( x = sort(sameFam_sameTAD_DT$dist), y = smooth_vals_sameFamSameTAD, col = sameFamSameTADcol)
lines( x = sort(sameFam_diffTAD_DT$dist), y = smooth_vals_sameFamDiffTAD, col = sameFamDiffTADcol)
legend("topright",
legend=c(paste0("sameFamSameTAD\n(AUC=", round(auc_sameFamSameTAD_obsDist, 2), ")"),
paste0("sameFamDiffTAD\n(AUC=", round(auc_sameFamDiffTAD_obsDist, 2))),
col = c(sameFamSameTADcol, sameFamDiffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
# PREDICT WITH DISTANCE VECTOR
smooth_vals_sameFamSameTAD_distVect <- predict(sameFamSameTAD_mod, distVect)
smooth_vals_sameFamDiffTAD_distVect <- predict(sameFamDiffTAD_mod, distVect)
auc_sameFamDiffTAD_distVect <- auc(x = distVect, y = smooth_vals_sameFamDiffTAD_distVect)
auc_sameFamSameTAD_distVect <- auc(x = distVect, y = smooth_vals_sameFamSameTAD_distVect)
outFile <- file.path(outFold, "diffTAD_mod.Rdata")
save(diffTAD_mod, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "sameFamDiffTAD_mod.Rdata")
save(sameFamDiffTAD_mod, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "sameTAD_mod.Rdata")
save(sameTAD_mod, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "sameFamSameTAD_mod.Rdata")
save(sameFamSameTAD_mod, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
sameFamDiffTAD_obsDist <- sameFam_diffTAD_DT$dist
outFile <- file.path(outFold, "sameFamDiffTAD_obsDist.Rdata")
save(sameFamDiffTAD_obsDist, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
sameFamSameTAD_obsDist <- sameFam_sameTAD_DT$dist
outFile <- file.path(outFold, "sameFamSameTAD_obsDist.Rdata")
save(sameFamSameTAD_obsDist, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "auc_sameFamDiffTAD_distVect.Rdata")
save(auc_sameFamDiffTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "auc_sameFamSameTAD_distVect.Rdata")
save(auc_sameFamSameTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "smooth_vals_sameFamSameTAD_distVect.Rdata")
save(smooth_vals_sameFamSameTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, "smooth_vals_sameFamDiffTAD_distVect.Rdata")
save(smooth_vals_sameFamDiffTAD_distVect, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameFamSameTAD_sameFamDiffTAD_loessFit_vectDist.", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(distVect),
ylim = range(c(na.omit(smooth_vals_sameFamSameTAD_distVect), na.omit(smooth_vals_sameFamDiffTAD_distVect))),
# xlab="",
# ylab="",
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = paste0("distance values seq from 0 to ", distLimit, " (# points = ", nbrLoessPoints, ")"), side = 3)
lines( x = distVect, y = smooth_vals_sameFamSameTAD_distVect, col = sameFamSameTADcol)
lines( x = distVect, y = smooth_vals_sameFamDiffTAD_distVect, col = sameFamDiffTADcol)
legend("topright",
legend=c(paste0("sameFamSameTAD\n(AUC=", round(auc_sameFamSameTAD_distVect, 2), ")"),
paste0("sameFamDiffTAD\n(AUC=", round(auc_sameFamDiffTAD_distVect, 2))),
col = c(sameFamSameTADcol, sameFamDiffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
################################################################
outFile <- file.path(outFold, paste0(curr_TADlist, "_", curr_dataset, "_sameTAD_diffTAD_sameFamSameTAD_sameFamDiffTAD_loessFit_vectDist.", plotType))
do.call(plotType, list(outFile, height = myHeight, width = myWidth))
plot(NULL,
xlim = range(distVect),
ylim = range(c(na.omit(smooth_vals_sameTAD_distVect),
na.omit(smooth_vals_sameFamSameTAD_distVect),
na.omit(smooth_vals_diffTAD_distVect),
na.omit(smooth_vals_sameFamDiffTAD_distVect))),
# xlab="",
# ylab="",
xlab=my_xlab,
ylab=my_ylab,
main=paste0(curr_TADlist, " - ", curr_dataset, ": coexpr ~ dist loess fit"))
mtext(text = paste0("distance values seq from 0 to ", distLimit, " (# points = ", nbrLoessPoints, ")"), side = 3)
lines( x = distVect, y = smooth_vals_sameFamSameTAD_distVect, col = sameFamSameTADcol)
lines( x = distVect, y = smooth_vals_sameFamDiffTAD_distVect, col = sameFamDiffTADcol)
lines( x = distVect, y = smooth_vals_sameTAD_distVect, col = sameTADcol)
lines( x = distVect, y = smooth_vals_diffTAD_distVect, col = diffTADcol)
legend("topright",
legend=c(paste0("sameTAD\n(AUC=", round(auc_sameTAD_distVect, 2), ")"),
paste0("diffTAD\n(AUC=", round(auc_diffTAD_distVect, 2)),
paste0("sameFamSameTAD\n(AUC=", round(auc_sameFamSameTAD_distVect, 2), ")"),
paste0("sameFamDiffTAD\n(AUC=", round(auc_sameFamDiffTAD_distVect, 2))),
col = c(sameTADcol, diffTADcol,
sameFamSameTADcol, sameFamDiffTADcol),
lty=1,
bty = "n")
foo <- dev.off()
cat(paste0("... written: ", outFile, "\n"))
################################################################
auc_values <- list(
auc_diffTAD_distVect = auc_diffTAD_distVect,
auc_sameTAD_distVect = auc_sameTAD_distVect,
auc_ratio_same_over_diff_distVect = auc_sameTAD_distVect/auc_diffTAD_distVect,
auc_diffTAD_obsDist = auc_diffTAD_obsDist,
auc_sameTAD_obsDist = auc_sameTAD_obsDist,
auc_ratio_same_over_diff_obsDist = auc_sameTAD_distVect/auc_diffTAD_obsDist,
auc_sameFamDiffTAD_distVect = auc_sameFamDiffTAD_distVect,
auc_sameFamSameTAD_distVect = auc_sameFamSameTAD_distVect,
auc_ratio_sameFam_same_over_diff_distVect = auc_sameFamSameTAD_distVect/auc_sameFamDiffTAD_distVect,
auc_sameFamDiffTAD_obsDist = auc_sameFamDiffTAD_obsDist,
auc_sameFamSameTAD_obsDist = auc_sameFamSameTAD_obsDist,
auc_ratio_sameFam_same_over_diff_obsDist = auc_sameFamSameTAD_distVect/auc_sameFamDiffTAD_obsDist
)
outFile <- file.path(outFold, paste0("auc_values.Rdata"))
save(auc_values, file = outFile)
cat(paste0("... written: ", outFile, "\n"))
} else{
stop("only loess implemented yet\n")
}
} # end iterating over family data
######################################################################################
######################################################################################
######################################################################################
cat(paste0("... written: ", logFile, "\n"))
######################################################################################
cat("*** DONE\n")
cat(paste0(startTime, "\n", Sys.time(), "\n"))
|
context("Screen flow data")
test_that("creates a dataframe with the proper columns", {
skip_on_cran()
skip_on_ci()
data <- screen_flow_data(station_number = "08NM116", include_symbols = FALSE)
expect_true(is.data.frame(data) &
ncol(data) == 22 &
all(c("Year","n_days","n_Q","n_missing_Q","Minimum","Jan_missing_Q") %in% colnames(data)))
})
test_that("outputs data for two stations", {
skip_on_cran()
skip_on_ci()
data <- screen_flow_data(station_number = c("08NM116","08HB048"), include_symbols = FALSE)
expect_true(length(unique(data$STATION_NUMBER)) == 2)
})
test_that("data is filtered by years properly", {
skip_on_cran()
skip_on_ci()
data <- screen_flow_data(station_number = "08NM116",
start_year = 1981,
end_year = 2010,
include_symbols = FALSE)
expect_identical(1981, min(data$Year))
expect_identical(2010, max(data$Year))
})
test_that("data is summarized by water years properly", {
skip_on_cran()
skip_on_ci()
flow_data <- add_date_variables(station_number = "08NM116",
water_year_start = 10)
test_data <- dplyr::filter(flow_data, WaterYear == 1981)
test_data <- dplyr::summarise(test_data,
Mean = mean(Value),
Median = median(Value),
Maximum = max(Value),
Minimum = min(Value))
data <- screen_flow_data(data = flow_data,
start_year = 1981,
water_year_start = 10,
include_symbols = FALSE)
data <- dplyr::filter(data, Year == 1981)
data <- dplyr::select(data, Mean, Median, Maximum, Minimum)
expect_equal(test_data, data)
})
test_that("missing dates are calculated properly", {
skip_on_cran()
skip_on_ci()
flow_data <- fill_missing_dates(station_number = "08NM116")
flow_data <- add_date_variables(flow_data)
test_data <- dplyr::summarise(dplyr::group_by(flow_data, WaterYear, MonthName),
sum = sum(is.na(Value)))
test_data <- tidyr::spread(test_data, MonthName, sum)
data <- screen_flow_data(station_number = "08NM116", include_symbols = FALSE)
expect_equal(test_data$Jan, data$Jan_missing_Q)
})
test_that("data is filtered by months properly", {
skip_on_cran()
skip_on_ci()
flow_data <- add_date_variables(station_number = "08NM116")
test_data <- dplyr::filter(flow_data,
WaterYear == 1981,
Month %in% 7:9)
test_data <- dplyr::summarise(test_data,
Mean = mean(Value),
Median = median(Value),
Maximum = max(Value),
Minimum = min(Value))
data <- screen_flow_data(data = flow_data,
start_year = 1981,
months = 7:9,
include_symbols = FALSE)
data_test <- dplyr::filter(data, Year == 1981)
data_test <- dplyr::select(data_test, Mean, Median, Maximum, Minimum)
expect_equal(test_data, data_test)
expect_true(ncol(data) == 13)
})
test_that("transpose properly transposed the results", {
skip_on_cran()
skip_on_ci()
data <- screen_flow_data(station_number = "08NM116",
transpose = TRUE, include_symbols = FALSE)
expect_true(all(c("n_days","n_Q","n_missing_Q","Minimum","Jan_missing_Q") %in% data$Statistic))
})
| /tests/testthat/test-screen_flow_data.R | no_license | cran/fasstr | R | false | false | 3,714 | r | context("Screen flow data")
test_that("creates a dataframe with the proper columns", {
skip_on_cran()
skip_on_ci()
data <- screen_flow_data(station_number = "08NM116", include_symbols = FALSE)
expect_true(is.data.frame(data) &
ncol(data) == 22 &
all(c("Year","n_days","n_Q","n_missing_Q","Minimum","Jan_missing_Q") %in% colnames(data)))
})
test_that("outputs data for two stations", {
skip_on_cran()
skip_on_ci()
data <- screen_flow_data(station_number = c("08NM116","08HB048"), include_symbols = FALSE)
expect_true(length(unique(data$STATION_NUMBER)) == 2)
})
test_that("data is filtered by years properly", {
skip_on_cran()
skip_on_ci()
data <- screen_flow_data(station_number = "08NM116",
start_year = 1981,
end_year = 2010,
include_symbols = FALSE)
expect_identical(1981, min(data$Year))
expect_identical(2010, max(data$Year))
})
test_that("data is summarized by water years properly", {
skip_on_cran()
skip_on_ci()
flow_data <- add_date_variables(station_number = "08NM116",
water_year_start = 10)
test_data <- dplyr::filter(flow_data, WaterYear == 1981)
test_data <- dplyr::summarise(test_data,
Mean = mean(Value),
Median = median(Value),
Maximum = max(Value),
Minimum = min(Value))
data <- screen_flow_data(data = flow_data,
start_year = 1981,
water_year_start = 10,
include_symbols = FALSE)
data <- dplyr::filter(data, Year == 1981)
data <- dplyr::select(data, Mean, Median, Maximum, Minimum)
expect_equal(test_data, data)
})
test_that("missing dates are calculated properly", {
skip_on_cran()
skip_on_ci()
flow_data <- fill_missing_dates(station_number = "08NM116")
flow_data <- add_date_variables(flow_data)
test_data <- dplyr::summarise(dplyr::group_by(flow_data, WaterYear, MonthName),
sum = sum(is.na(Value)))
test_data <- tidyr::spread(test_data, MonthName, sum)
data <- screen_flow_data(station_number = "08NM116", include_symbols = FALSE)
expect_equal(test_data$Jan, data$Jan_missing_Q)
})
test_that("data is filtered by months properly", {
skip_on_cran()
skip_on_ci()
flow_data <- add_date_variables(station_number = "08NM116")
test_data <- dplyr::filter(flow_data,
WaterYear == 1981,
Month %in% 7:9)
test_data <- dplyr::summarise(test_data,
Mean = mean(Value),
Median = median(Value),
Maximum = max(Value),
Minimum = min(Value))
data <- screen_flow_data(data = flow_data,
start_year = 1981,
months = 7:9,
include_symbols = FALSE)
data_test <- dplyr::filter(data, Year == 1981)
data_test <- dplyr::select(data_test, Mean, Median, Maximum, Minimum)
expect_equal(test_data, data_test)
expect_true(ncol(data) == 13)
})
test_that("transpose properly transposed the results", {
skip_on_cran()
skip_on_ci()
data <- screen_flow_data(station_number = "08NM116",
transpose = TRUE, include_symbols = FALSE)
expect_true(all(c("n_days","n_Q","n_missing_Q","Minimum","Jan_missing_Q") %in% data$Statistic))
})
|
> ###Random Forest top 5 variables based on importance: strikeouts, wins, earned run average, at bats, and innings pitched
> ###Find players who have same minimum amount of strikeouts, wins, earned run average, at bats, and innings pitched as award winners
> Pitching_Stats_Full_Dataset <- read_csv("E:/Grad School/Practicum 2/Pitching Stats Full Dataset.csv",
+ col_types = cols(`2b ` = col_number(),
+ `3b ` = col_number(), Award = col_number(),
+ ConfStanding = col_number(), Year = col_number(),
+ `ab ` = col_number(), app = col_number(),
+ `bavg ` = col_number(), `bb ` = col_number(),
+ `bk ` = col_number(), `cg ` = col_number(),
+ csho = col_number(), `er ` = col_number(),
+ `era ` = col_number(), `gs ` = col_number(),
+ `h ` = col_number(), `hbp ` = col_number(),
+ `hr ` = col_number(), `ip ` = col_number(),
+ `isho ` = col_number(), `l ` = col_number(),
+ `r ` = col_number(), `so ` = col_number(),
+ `sv ` = col_number(), w = col_number(),
+ `wp ` = col_number()), locale = locale(encoding = "ASCII"))
> psfd <- Pitching_Stats_Full_Dataset
> library(dplyr)
> first <- filter(psfd, psfd$Award == 1)
> second <- filter(psfd, psfd$Award == 2)
> noteam <- filter(psfd, psfd$Award == 0)
> summary(psfd)
> psfd$Award <- as.factor(psfd$Award)
> summary(first)
> ###Stat Requirements needed for 1st Team: strikeouts >= 45, wins >= 6, earned run average <= 4.287, at bats >= 228, innings pitched >= 60.83
> firstteamcandidates <- filter(psfd, psfd$Award == 0 | psfd$Award == 2)
> ftc_stats <- filter(firstteamcandidates, firstteamcandidates$`so ` >= 45 & firstteamcandidates$w >= 6 & firstteamcandidates$`era ` <= 4.287 & firstteamcandidates$`ab ` >= 228 & firstteamcandidates$`ip ` >= 60.83)
> View(ftc_stats)
> write.csv(ftc_stats, "First Team Candidates.csv") | /Pitching Candidates.R | no_license | dlacomb/RMACPitchingVoting | R | false | false | 2,717 | r | > ###Random Forest top 5 variables based on importance: strikeouts, wins, earned run average, at bats, and innings pitched
> ###Find players who have same minimum amount of strikeouts, wins, earned run average, at bats, and innings pitched as award winners
> Pitching_Stats_Full_Dataset <- read_csv("E:/Grad School/Practicum 2/Pitching Stats Full Dataset.csv",
+ col_types = cols(`2b ` = col_number(),
+ `3b ` = col_number(), Award = col_number(),
+ ConfStanding = col_number(), Year = col_number(),
+ `ab ` = col_number(), app = col_number(),
+ `bavg ` = col_number(), `bb ` = col_number(),
+ `bk ` = col_number(), `cg ` = col_number(),
+ csho = col_number(), `er ` = col_number(),
+ `era ` = col_number(), `gs ` = col_number(),
+ `h ` = col_number(), `hbp ` = col_number(),
+ `hr ` = col_number(), `ip ` = col_number(),
+ `isho ` = col_number(), `l ` = col_number(),
+ `r ` = col_number(), `so ` = col_number(),
+ `sv ` = col_number(), w = col_number(),
+ `wp ` = col_number()), locale = locale(encoding = "ASCII"))
> psfd <- Pitching_Stats_Full_Dataset
> library(dplyr)
> first <- filter(psfd, psfd$Award == 1)
> second <- filter(psfd, psfd$Award == 2)
> noteam <- filter(psfd, psfd$Award == 0)
> summary(psfd)
> psfd$Award <- as.factor(psfd$Award)
> summary(first)
> ###Stat Requirements needed for 1st Team: strikeouts >= 45, wins >= 6, earned run average <= 4.287, at bats >= 228, innings pitched >= 60.83
> firstteamcandidates <- filter(psfd, psfd$Award == 0 | psfd$Award == 2)
> ftc_stats <- filter(firstteamcandidates, firstteamcandidates$`so ` >= 45 & firstteamcandidates$w >= 6 & firstteamcandidates$`era ` <= 4.287 & firstteamcandidates$`ab ` >= 228 & firstteamcandidates$`ip ` >= 60.83)
> View(ftc_stats)
> write.csv(ftc_stats, "First Team Candidates.csv") |
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/predict_mbsts.R
\name{predict.mbsts}
\alias{predict.mbsts}
\title{Predictions for a given multivariate Bayesian structural time series model}
\usage{
\method{predict}{mbsts}(object, steps.ahead, newdata = NULL, ...)
}
\arguments{
\item{object}{An object of class 'mbsts', a result of call to \code{\link{as.mbsts}}.}
\item{steps.ahead}{An integer value (S) specifying the number of steps ahead
to be forecasted. If 'object' contains a regression component, the argument
is disregarded and a prediction is made with the same length of
'newdata'.}
\item{newdata}{Optional matrix of new data. Only required when 'object'
contains a regression component.}
\item{...}{Arguments passed to other methods (currently unused).}
}
\value{
Returns a list with the following components
\describe{
\item{post.pred.0}{t x d x ('niter'- 'burn') array of in-sample forecasts.}
\item{post.pred.1}{S x d x ('niter'- 'burn') array out-of-sample forecasts, where S is
the number of forecasted periods (set to the length of 'steps.ahead' or 'newdata').}
\item{post.pred}{(t + S) x d x ('niter'- 'burn') array combining in- and out-of-sample forecasts.}
}
}
\description{
Given an object of class 'mbsts' and the number of 'steps.ahead' in the future to be
forecasted, this function provides in-sample forecasts and out-of-sample forecasts,
both based on drawing from the posterior predictive distribution. If 'mbsts' contains a
regression component, then the new matrix of predictors 'newdata' must be provided.
Note that NA values are not allowed in the new regressor matrix.
}
\examples{
## Note: The following are toy examples, for a proper analysis we recommend to run
## at least 1000 iterations and check the convergence of the Markov chain
## Example 1 :
y <- cbind(seq(0.5,100,by=0.5)*0.1 + rnorm(200),
seq(100.25,150,by=0.25)*0.05 + rnorm(200),
rnorm(200, 5,1))
mbsts.1 <- as.mbsts(y = y, components = c("trend", "seasonal"), seas.period = 7, s0.r = diag(3),
s0.eps = diag(3), niter = 50, burn = 5)
pred.1 <- predict(mbsts.1, steps.ahead = 10)
## Example 2
y <- cbind(rnorm(100), rnorm(100, 2, 3))
X <- cbind(rnorm(100, 0.5, 1) + 5, rnorm(100, 0.2, 2) - 2)
mbsts.2 <- as.mbsts(y = y, components = c("trend", "seasonal"),
seas.period = 7, X = X, s0.r = diag(2),
s0.eps = diag(2), niter = 100, burn = 10)
newdata <- cbind(rnorm(30), rt(30, 2))
pred.2 <- predict(mbsts.2, newdata = newdata)
}
| /man/predict.mbsts.Rd | no_license | FMenchetti/CausalMBSTS | R | false | true | 2,562 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/predict_mbsts.R
\name{predict.mbsts}
\alias{predict.mbsts}
\title{Predictions for a given multivariate Bayesian structural time series model}
\usage{
\method{predict}{mbsts}(object, steps.ahead, newdata = NULL, ...)
}
\arguments{
\item{object}{An object of class 'mbsts', a result of call to \code{\link{as.mbsts}}.}
\item{steps.ahead}{An integer value (S) specifying the number of steps ahead
to be forecasted. If 'object' contains a regression component, the argument
is disregarded and a prediction is made with the same length of
'newdata'.}
\item{newdata}{Optional matrix of new data. Only required when 'object'
contains a regression component.}
\item{...}{Arguments passed to other methods (currently unused).}
}
\value{
Returns a list with the following components
\describe{
\item{post.pred.0}{t x d x ('niter'- 'burn') array of in-sample forecasts.}
\item{post.pred.1}{S x d x ('niter'- 'burn') array out-of-sample forecasts, where S is
the number of forecasted periods (set to the length of 'steps.ahead' or 'newdata').}
\item{post.pred}{(t + S) x d x ('niter'- 'burn') array combining in- and out-of-sample forecasts.}
}
}
\description{
Given an object of class 'mbsts' and the number of 'steps.ahead' in the future to be
forecasted, this function provides in-sample forecasts and out-of-sample forecasts,
both based on drawing from the posterior predictive distribution. If 'mbsts' contains a
regression component, then the new matrix of predictors 'newdata' must be provided.
Note that NA values are not allowed in the new regressor matrix.
}
\examples{
## Note: The following are toy examples, for a proper analysis we recommend to run
## at least 1000 iterations and check the convergence of the Markov chain
## Example 1 :
y <- cbind(seq(0.5,100,by=0.5)*0.1 + rnorm(200),
seq(100.25,150,by=0.25)*0.05 + rnorm(200),
rnorm(200, 5,1))
mbsts.1 <- as.mbsts(y = y, components = c("trend", "seasonal"), seas.period = 7, s0.r = diag(3),
s0.eps = diag(3), niter = 50, burn = 5)
pred.1 <- predict(mbsts.1, steps.ahead = 10)
## Example 2
y <- cbind(rnorm(100), rnorm(100, 2, 3))
X <- cbind(rnorm(100, 0.5, 1) + 5, rnorm(100, 0.2, 2) - 2)
mbsts.2 <- as.mbsts(y = y, components = c("trend", "seasonal"),
seas.period = 7, X = X, s0.r = diag(2),
s0.eps = diag(2), niter = 100, burn = 10)
newdata <- cbind(rnorm(30), rt(30, 2))
pred.2 <- predict(mbsts.2, newdata = newdata)
}
|
#' timeFloor
#'
#' This function acts like a floor function execpt for time.
#'
#' @param timeValue typically a column of time Values to take the floor function
#' @param timeIntervalValue the value of the time interval
#' @param timeIntervalUnits the units of the time interval
#' @keywords time
#' @examples
#'
#' library(EmissionsHelper)
#' mytime <- reformatTime(c("2018-1-1 12:01", "2018-1-1 12:02",
#' "2018-1-1 12:16", "2018-1-1 12:17"))
#' timeFloor(mytime, 15, "minutes")
#'
#' @export
timeFloor <- function(timeValue, timeIntervalValue = 15, timeIntervalUnits = "minutes"){
timeIntervalInSeconds <- switch(timeIntervalUnits,
"seconds" = 1,
"second" = 1,
"sec" = 1,
"s" = 1,
"minutes" = 60,
"minute" = 60,
"min" = 60,
"m" = 60,
"hours" = 3600,
"hour" = 3600,
"hr" = 3600,
"h" = 3600)
if(!is.null(timeValue)){
return(as.POSIXct(floor(as.numeric(as.POSIXct(timeValue)) /
(timeIntervalValue * timeIntervalInSeconds)) *
(timeIntervalValue * timeIntervalInSeconds),
origin='1970-01-01'))
}else{
warning("Check parameters in timeFloor function ... timeValue is NULL")
}
}
| /R/timeFloor.R | permissive | JerryHMartin/EmissionsHelper | R | false | false | 1,661 | r | #' timeFloor
#'
#' This function acts like a floor function execpt for time.
#'
#' @param timeValue typically a column of time Values to take the floor function
#' @param timeIntervalValue the value of the time interval
#' @param timeIntervalUnits the units of the time interval
#' @keywords time
#' @examples
#'
#' library(EmissionsHelper)
#' mytime <- reformatTime(c("2018-1-1 12:01", "2018-1-1 12:02",
#' "2018-1-1 12:16", "2018-1-1 12:17"))
#' timeFloor(mytime, 15, "minutes")
#'
#' @export
timeFloor <- function(timeValue, timeIntervalValue = 15, timeIntervalUnits = "minutes"){
timeIntervalInSeconds <- switch(timeIntervalUnits,
"seconds" = 1,
"second" = 1,
"sec" = 1,
"s" = 1,
"minutes" = 60,
"minute" = 60,
"min" = 60,
"m" = 60,
"hours" = 3600,
"hour" = 3600,
"hr" = 3600,
"h" = 3600)
if(!is.null(timeValue)){
return(as.POSIXct(floor(as.numeric(as.POSIXct(timeValue)) /
(timeIntervalValue * timeIntervalInSeconds)) *
(timeIntervalValue * timeIntervalInSeconds),
origin='1970-01-01'))
}else{
warning("Check parameters in timeFloor function ... timeValue is NULL")
}
}
|
AL.Batting=read.table("AL.Batting.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Pitching=read.table("AL.Pitching.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Fielding=read.table("AL.Fielding.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Batting=read.table("NL.Batting.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Pitching=read.table("NL.Pitching.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Fielding=read.table("NL.Fielding.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Batting.2014=read.table("AL.Batting.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Pitching.2014=read.table("AL.Pitching.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Fielding.2014=read.table("AL.Fielding.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Batting.2014=read.table("NL.Batting.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Pitching.2014=read.table("NL.Pitching.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Fielding.2014=read.table("NL.Fielding.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL=cbind(AL.Batting,AL.Pitching[,-1],AL.Fielding[,-1])
NL=cbind(NL.Batting,NL.Pitching[,-1],NL.Fielding[,-1])
AL.year=rep(c("09","10","11","12","13"),c(14,14,14,14,15))
NL.year=rep(c("09","10","11","12","13"),c(16,16,16,16,15))
AL$Tm=paste(AL.year,AL$Tm)
NL$Tm=paste(NL.year,NL$Tm)
AL.2014=cbind(AL.Batting.2014,AL.Pitching.2014[,-1],AL.Fielding.2014[,-1])
NL.2014=cbind(NL.Batting.2014,NL.Pitching.2014[,-1],NL.Fielding.2014[,-1])
AL.2014$Tm=paste(rep("14",15),AL.2014$Tm)
NL.2014$Tm=paste(rep("14",15),NL.2014$Tm)
AL.TOTAL=rbind(AL,AL.2014)
NL.TOTAL=rbind(NL,NL.2014)
##############################################################
AL.model=kmeans(AL.TOTAL[,-1],10,nstart=length(AL.TOTAL$Tm))
AL.cluster.team=data.frame(AL.TOTAL$Tm,AL.model$cluster)
NL.model=kmeans(NL.TOTAL[,-1],10,nstart=length(NL.TOTAL$Tm))
NL.cluster.team=data.frame(NL.TOTAL$Tm,10+NL.model$cluster)
##################################################################
library(XML)
library(RCurl)
webpage<-getURL("http://www.baseball-reference.com/teams/STL/2009-schedule-scores.shtml")
webpage <- readLines(tc <- textConnection(webpage)); close(tc)
pagetree <- htmlTreeParse(webpage, error=function(...){}, useInternalNodes = TRUE)
tablelines <- xpathSApply(pagetree, "//tr[@class='']", xmlValue)
n=length(tablelines)
lines=strsplit(tablelines[2:n],"\n ",fixed=FALSE, useBytes = TRUE)
datamatrix=matrix(0,length(lines),6)
for(i in 1:length(lines))
{
datamatrix[i,1:6]=(lines[[i]])[c(5:7,9:10,21)]
}
datamatrix[,1]=sub(" ",replacement="",datamatrix[,1])
datamatrix[,2]=ifelse(datamatrix[,2]==" @","Guest","Home")
datamatrix[,3]=sub(" ",replacement="",datamatrix[,3])
datamatrix[,4]=sub(" ",replacement="",datamatrix[,4])
datamatrix[,4]=sub("Game Preview, Matchups, and Tickets",replacement="",datamatrix[,4])
datamatrix[,5]=sub(" ",replacement="",datamatrix[,5])
datamatrix[,5]=sub("\n",replacement="",datamatrix[,5])
datamatrix[,6]=sub("\n",replacement="",datamatrix[,6])
datamatrix[,6]=sub(" ",replacement="",datamatrix[,6])
datamatrix=datamatrix[!(datamatrix[,4]==""),]
tidydata=matrix(0,length(datamatrix[,1]),5)
for(i in 1:length(datamatrix[,1]))
{
tidydata[i,1]=ifelse(datamatrix[i,2]=="Guest",datamatrix[i,1],datamatrix[i,3])
tidydata[i,2]=ifelse(datamatrix[i,2]=="Home",datamatrix[i,1],datamatrix[i,3])
tidydata[i,3]=ifelse(datamatrix[i,2]=="Guest",datamatrix[i,4],datamatrix[i,5])
tidydata[i,4]=ifelse(datamatrix[i,2]=="Home",datamatrix[i,4],datamatrix[i,5])
}
| /mlb.R | no_license | ShaneKao/SportsLottery | R | false | false | 3,560 | r | AL.Batting=read.table("AL.Batting.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Pitching=read.table("AL.Pitching.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Fielding=read.table("AL.Fielding.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Batting=read.table("NL.Batting.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Pitching=read.table("NL.Pitching.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Fielding=read.table("NL.Fielding.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Batting.2014=read.table("AL.Batting.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Pitching.2014=read.table("AL.Pitching.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL.Fielding.2014=read.table("AL.Fielding.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Batting.2014=read.table("NL.Batting.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Pitching.2014=read.table("NL.Pitching.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
NL.Fielding.2014=read.table("NL.Fielding.2014.txt",sep=",",header=TRUE,stringsAsFactors=FALSE)
AL=cbind(AL.Batting,AL.Pitching[,-1],AL.Fielding[,-1])
NL=cbind(NL.Batting,NL.Pitching[,-1],NL.Fielding[,-1])
AL.year=rep(c("09","10","11","12","13"),c(14,14,14,14,15))
NL.year=rep(c("09","10","11","12","13"),c(16,16,16,16,15))
AL$Tm=paste(AL.year,AL$Tm)
NL$Tm=paste(NL.year,NL$Tm)
AL.2014=cbind(AL.Batting.2014,AL.Pitching.2014[,-1],AL.Fielding.2014[,-1])
NL.2014=cbind(NL.Batting.2014,NL.Pitching.2014[,-1],NL.Fielding.2014[,-1])
AL.2014$Tm=paste(rep("14",15),AL.2014$Tm)
NL.2014$Tm=paste(rep("14",15),NL.2014$Tm)
AL.TOTAL=rbind(AL,AL.2014)
NL.TOTAL=rbind(NL,NL.2014)
##############################################################
AL.model=kmeans(AL.TOTAL[,-1],10,nstart=length(AL.TOTAL$Tm))
AL.cluster.team=data.frame(AL.TOTAL$Tm,AL.model$cluster)
NL.model=kmeans(NL.TOTAL[,-1],10,nstart=length(NL.TOTAL$Tm))
NL.cluster.team=data.frame(NL.TOTAL$Tm,10+NL.model$cluster)
##################################################################
library(XML)
library(RCurl)
webpage<-getURL("http://www.baseball-reference.com/teams/STL/2009-schedule-scores.shtml")
webpage <- readLines(tc <- textConnection(webpage)); close(tc)
pagetree <- htmlTreeParse(webpage, error=function(...){}, useInternalNodes = TRUE)
tablelines <- xpathSApply(pagetree, "//tr[@class='']", xmlValue)
n=length(tablelines)
lines=strsplit(tablelines[2:n],"\n ",fixed=FALSE, useBytes = TRUE)
datamatrix=matrix(0,length(lines),6)
for(i in 1:length(lines))
{
datamatrix[i,1:6]=(lines[[i]])[c(5:7,9:10,21)]
}
datamatrix[,1]=sub(" ",replacement="",datamatrix[,1])
datamatrix[,2]=ifelse(datamatrix[,2]==" @","Guest","Home")
datamatrix[,3]=sub(" ",replacement="",datamatrix[,3])
datamatrix[,4]=sub(" ",replacement="",datamatrix[,4])
datamatrix[,4]=sub("Game Preview, Matchups, and Tickets",replacement="",datamatrix[,4])
datamatrix[,5]=sub(" ",replacement="",datamatrix[,5])
datamatrix[,5]=sub("\n",replacement="",datamatrix[,5])
datamatrix[,6]=sub("\n",replacement="",datamatrix[,6])
datamatrix[,6]=sub(" ",replacement="",datamatrix[,6])
datamatrix=datamatrix[!(datamatrix[,4]==""),]
tidydata=matrix(0,length(datamatrix[,1]),5)
for(i in 1:length(datamatrix[,1]))
{
tidydata[i,1]=ifelse(datamatrix[i,2]=="Guest",datamatrix[i,1],datamatrix[i,3])
tidydata[i,2]=ifelse(datamatrix[i,2]=="Home",datamatrix[i,1],datamatrix[i,3])
tidydata[i,3]=ifelse(datamatrix[i,2]=="Guest",datamatrix[i,4],datamatrix[i,5])
tidydata[i,4]=ifelse(datamatrix[i,2]=="Home",datamatrix[i,4],datamatrix[i,5])
}
|
#' BUS
#'
#' The class BUS
#'
#' @field n total number of buses
#' @field index bus indexes
#' @field int internal nus number
#' @field Pg active power injected in the network by generators
#' @field Qg reactive power injected in the network by generators
#' @field Pl active power absorbed from the network by loads
#' @field Ql reactive power absorbed from the network by loads
#' @field names bus names
BUS <- setRefClass("BUS", contains = "powerr",
fields = list(n = "numeric",
index = "numeric",
int = "numeric",
a = "numeric",
v = "numeric",
Pg = "numeric",
Qg = "numeric",
Pl = "numeric",
Ql = "numeric",
names = "character"),
methods = list(
initialize = function(data, n, index, int, a, v, Pg, Qg, Pl, Ql, names, ncol){
n <<- numeric();
a <<- numeric();
v <<- numeric();
index <<- numeric();
int <<- numeric();
Pg <<- numeric();
Qg <<- numeric();
Pl <<- numeric();
Ql <<- numeric();
names <<- character();
ncol <<- 6;
},
setup = function(){
if (nrow(data) == 0 && ncol(data) == 0){
print('The data file does not seem to be in a valid format: no bus found$')
}
n <<- nrow(data);
a <<- 1:n;
v <<- a + n;
# setup internal bus number for second indexing of bus
int[data[, 1]] <<- a;
.GlobalEnv$DAE$m <- 2 * n;
.GlobalEnv$DAE$y <- rep(0, .GlobalEnv$DAE$m);
.GlobalEnv$DAE$g <- rep(0, .GlobalEnv$DAE$m);
# .GlobalEnv$DAE$Gy <- Matrix(0, nrow = .GlobalEnv$DAE$m, ncol = .GlobalEnv$DAE$m, sparse = TRUE);
.GlobalEnv$DAE$Gy <- matrix(0, nrow = .GlobalEnv$DAE$m, ncol = .GlobalEnv$DAE$m);
if (ncol(data) >= 4){
# check voltage magnitudes
if (sum(data[, 3] < 0.5) > 0){
warning('some initial guess voltage magnitudes are too low.');
}
if (sum(data[, 3] > 1.5) > 0){
warning('some initial guess voltage magnitudes are too high.');
}
.GlobalEnv$DAE$y[v] <- data[, 3];
# check voltage phases
if (sum(data[, 4] < -1.5708) > 0){
warning('some initial guess voltage phases are too low.');
}
if (sum(data[, 4] > 1.5708) > 0){
warning('some initial guess voltage phases are too high.');
}
.GlobalEnv$DAE$y[a] <- data[, 4];
}
else{
.GlobalEnv$DAE$y[a] <- rep(0, n);
.GlobalEnv$DAE$y[v] <- rep(1, n);
}
Pl <<- rep(0, n);
Ql <<- rep(0, n);
Pg <<- rep(0, n);
Qg <<- rep(0, n);
},
getbus = function(idx){
uTemp <- int[round(idx)];
vTemp <- uTemp + n;
return(list(uTemp, vTemp));
},
getzeros = function(){
uTemp <- rep(0, n);
return(uTemp);
},
getint = function(idx){
uTemp <- int[round(idx)];
return(uTemp);
},
test = function(){
print(environment())
print(parent.env(environment()))
print(parent.env(parent.env(environment())))
print(parent.env(parent.env(parent.env(environment()))))
print(parent.env(parent.env(parent.env(parent.env(environment())))))
.GlobalEnv$DAE$m <- 1000;
.GlobalEnv$DAE$n <- .GlobalEnv$DAE$m;
print(.GlobalEnv$DAE$n)
}
)) | /R/class_BUS.R | no_license | Jingfan-Sun/powerr | R | false | false | 5,440 | r | #' BUS
#'
#' The class BUS
#'
#' @field n total number of buses
#' @field index bus indexes
#' @field int internal nus number
#' @field Pg active power injected in the network by generators
#' @field Qg reactive power injected in the network by generators
#' @field Pl active power absorbed from the network by loads
#' @field Ql reactive power absorbed from the network by loads
#' @field names bus names
BUS <- setRefClass("BUS", contains = "powerr",
fields = list(n = "numeric",
index = "numeric",
int = "numeric",
a = "numeric",
v = "numeric",
Pg = "numeric",
Qg = "numeric",
Pl = "numeric",
Ql = "numeric",
names = "character"),
methods = list(
initialize = function(data, n, index, int, a, v, Pg, Qg, Pl, Ql, names, ncol){
n <<- numeric();
a <<- numeric();
v <<- numeric();
index <<- numeric();
int <<- numeric();
Pg <<- numeric();
Qg <<- numeric();
Pl <<- numeric();
Ql <<- numeric();
names <<- character();
ncol <<- 6;
},
setup = function(){
if (nrow(data) == 0 && ncol(data) == 0){
print('The data file does not seem to be in a valid format: no bus found$')
}
n <<- nrow(data);
a <<- 1:n;
v <<- a + n;
# setup internal bus number for second indexing of bus
int[data[, 1]] <<- a;
.GlobalEnv$DAE$m <- 2 * n;
.GlobalEnv$DAE$y <- rep(0, .GlobalEnv$DAE$m);
.GlobalEnv$DAE$g <- rep(0, .GlobalEnv$DAE$m);
# .GlobalEnv$DAE$Gy <- Matrix(0, nrow = .GlobalEnv$DAE$m, ncol = .GlobalEnv$DAE$m, sparse = TRUE);
.GlobalEnv$DAE$Gy <- matrix(0, nrow = .GlobalEnv$DAE$m, ncol = .GlobalEnv$DAE$m);
if (ncol(data) >= 4){
# check voltage magnitudes
if (sum(data[, 3] < 0.5) > 0){
warning('some initial guess voltage magnitudes are too low.');
}
if (sum(data[, 3] > 1.5) > 0){
warning('some initial guess voltage magnitudes are too high.');
}
.GlobalEnv$DAE$y[v] <- data[, 3];
# check voltage phases
if (sum(data[, 4] < -1.5708) > 0){
warning('some initial guess voltage phases are too low.');
}
if (sum(data[, 4] > 1.5708) > 0){
warning('some initial guess voltage phases are too high.');
}
.GlobalEnv$DAE$y[a] <- data[, 4];
}
else{
.GlobalEnv$DAE$y[a] <- rep(0, n);
.GlobalEnv$DAE$y[v] <- rep(1, n);
}
Pl <<- rep(0, n);
Ql <<- rep(0, n);
Pg <<- rep(0, n);
Qg <<- rep(0, n);
},
getbus = function(idx){
uTemp <- int[round(idx)];
vTemp <- uTemp + n;
return(list(uTemp, vTemp));
},
getzeros = function(){
uTemp <- rep(0, n);
return(uTemp);
},
getint = function(idx){
uTemp <- int[round(idx)];
return(uTemp);
},
test = function(){
print(environment())
print(parent.env(environment()))
print(parent.env(parent.env(environment())))
print(parent.env(parent.env(parent.env(environment()))))
print(parent.env(parent.env(parent.env(parent.env(environment())))))
.GlobalEnv$DAE$m <- 1000;
.GlobalEnv$DAE$n <- .GlobalEnv$DAE$m;
print(.GlobalEnv$DAE$n)
}
)) |
dat <- sp500ret$sp500RET
ts.plot(dt)
acf(dat) # doesnt tell you to look at 2nd order
acf(dat^2)
McLeod.Li.test(y=dat) # want to reject the null hypothesis
McLeod.Li.test(y=mod$residuals)
# fGarch uses ljung box
# https://quant.stackexchange.com/questions/11019/garch-model-and-prediction
| /TimeSeries/Week8_Intervention_ARCH/McLeod_Li_test.R | no_license | jfnavarro21/MSCA_Python_R | R | false | false | 291 | r | dat <- sp500ret$sp500RET
ts.plot(dt)
acf(dat) # doesnt tell you to look at 2nd order
acf(dat^2)
McLeod.Li.test(y=dat) # want to reject the null hypothesis
McLeod.Li.test(y=mod$residuals)
# fGarch uses ljung box
# https://quant.stackexchange.com/questions/11019/garch-model-and-prediction
|
kernelFun <- function(type=c('truncated','bartlett','daniell','QS','parzen'),z){
type <- match.arg(type)
if (missing(type)) stop("Kernel type is missing. Type must be one of 'truncated','bartlett','daniell','QS','parzen'")
if (type != "truncated" && type != "bartlett" && type != "daniell" && type != "QS" && type != "parzen") stop("Wrong type. Type must be one of 'truncated','bartlett','daniell','QS','parzen'")
if (length(z)>1) stop("z must be of length 1")
if (!is.numeric(z))
stop("'z' must be numeric")
if (type=="truncated"){
k <- ifelse(abs(z)<=1,1,0)
}
else if (type=="bartlett"){
k <- ifelse(abs(z)<=1,1-abs(z),0)
}
else if (type=="daniell"){
if (z==0) stop("z cannot be zero")
k <- sin(pi*z)/(pi*z)
}
else if (type=="QS"){
k <- (9/(5*pi^2*z^2))*(sin(sqrt(5/3)*pi*z)/(sqrt(5/3)*pi*z) - cos(sqrt(5/3)*pi*z))
}
else if (type=="parzen"){
if (abs(z) <= (3/pi)){
k <- 1-6*((pi*z)/6)^2 + 6*abs((pi*z)/6)^3
}
else if ((abs(z) >= (3/pi)) && (abs(z) <= (6/pi))){
k <- 2*(1-abs((pi*z)/6))^3
}
else {
k <- 0
}
}
return(k)
}
| /dCovTS/R/kernelFun.R | no_license | ingted/R-Examples | R | false | false | 1,112 | r | kernelFun <- function(type=c('truncated','bartlett','daniell','QS','parzen'),z){
type <- match.arg(type)
if (missing(type)) stop("Kernel type is missing. Type must be one of 'truncated','bartlett','daniell','QS','parzen'")
if (type != "truncated" && type != "bartlett" && type != "daniell" && type != "QS" && type != "parzen") stop("Wrong type. Type must be one of 'truncated','bartlett','daniell','QS','parzen'")
if (length(z)>1) stop("z must be of length 1")
if (!is.numeric(z))
stop("'z' must be numeric")
if (type=="truncated"){
k <- ifelse(abs(z)<=1,1,0)
}
else if (type=="bartlett"){
k <- ifelse(abs(z)<=1,1-abs(z),0)
}
else if (type=="daniell"){
if (z==0) stop("z cannot be zero")
k <- sin(pi*z)/(pi*z)
}
else if (type=="QS"){
k <- (9/(5*pi^2*z^2))*(sin(sqrt(5/3)*pi*z)/(sqrt(5/3)*pi*z) - cos(sqrt(5/3)*pi*z))
}
else if (type=="parzen"){
if (abs(z) <= (3/pi)){
k <- 1-6*((pi*z)/6)^2 + 6*abs((pi*z)/6)^3
}
else if ((abs(z) >= (3/pi)) && (abs(z) <= (6/pi))){
k <- 2*(1-abs((pi*z)/6))^3
}
else {
k <- 0
}
}
return(k)
}
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/GraphFunctions.R
\name{isConnected.adjMatrixGraph}
\alias{isConnected.adjMatrixGraph}
\title{A Graph Function}
\usage{
\method{isConnected}{adjMatrixGraph}(graph)
}
\arguments{
\item{graph}{the graph obj}
}
\value{
a list containing the graph and whether the graph is connected
}
\description{
This functions checks if the graph is connected.
}
| /man/isConnected.adjMatrixGraph.Rd | no_license | TheBell/Graph | R | false | true | 440 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/GraphFunctions.R
\name{isConnected.adjMatrixGraph}
\alias{isConnected.adjMatrixGraph}
\title{A Graph Function}
\usage{
\method{isConnected}{adjMatrixGraph}(graph)
}
\arguments{
\item{graph}{the graph obj}
}
\value{
a list containing the graph and whether the graph is connected
}
\description{
This functions checks if the graph is connected.
}
|
source('simulation/sim_model.R')
#install.packages(c("MFPCA","funData"))
#install.packages("funData")
source('nafpca.R')
source('../../classification.R')
library(MFPCA);library(funData)
library(dplyr)
library(ggplot2)
scaling <- function(x,k) {
#x = abs(x)
out = (x - min(x)) / (max(x) - min(x)) / k
return(out)
}
##################################################################
# Model I-1 : 1-d functional data - nonlinear relation in clustering
##################################################################
n=200; nt=20
set.seed(1)
dat = sim.model.11(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
# Original Data
plot(tt,x[1,], type='l', ylim=c(-5,5), xlab="t", ylab="X(t)", main = "Model I", col=y+1)
for(i in 1:n)lines(tt,x[i,], col=(y[i]+1))
# function estimation
# 1. fda-based
# bspline: most natural
coef.bspline = get.fd(t(x),tt ,basisname='bspline')
fd.bsp = fd(coef.bspline$coef, coef.bspline$basis)
plot(fd.bsp, col=y+1)
evals = t(eval.fd(tt,fd.bsp))
sum((evals - x)^2) / n /nt
# fourier: tend to get x[0] = 0 , x[1] =0
coef.fourier = get.fd(t(x),tt ,basisname='fourier', nbasis=53)
fd.fourier=fd(coef.fourier$coef, coef.fourier$basis)
plot(fd.fourier, col=y+1)
evals = t(eval.fd(tt,fd.fourier))
sum((evals - x)^2) / n /nt
# 2. rkhs-based
# gaussian: natural
coef.gauss = get.fd.rkhs(x, tt, kern="gauss")
evals = coef.gauss$coef %*% coef.gauss$kt
plot(tt, evals[1,], col=y[1]+1, type='l', ylim=c(min(evals),max(evals)))
for(j in 1:n){
lines(tt, evals[j,], col=y[j]+1)
}
sum((evals - x)^2) / n /nt
# brownian: natural --- the best in this situation
coef.brown = get.fd.rkhs(x, tt, kern="brown")
evals = coef.brown$coef %*% coef.brown$kt
plot(tt, evals[1,], col=y[1]+1, type='l', ylim=c(min(evals),max(evals)))
for(j in 1:n){
lines(tt, evals[j,], col=y[j]+1)
}
sum((evals - x)^2) / n /nt
# Nonlinear FPCA
tmp=nafpca(x,tt, shx=11, nbasis=21,gamma.tune = TRUE,
basisname='bspline')
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='fourier')
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=30,
basisname='gauss')
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=20,
basisname='brown')
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=20,
basisname='bspline', kernel='poly',
c=.1,d=2)
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='fourier', kernel='poly',
c=.5,d=2)
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='gauss', kernel='poly',
c=.5,d=2)
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='brown', kernel='poly',
c=.5,d=2)
tmp$eval
gk = Gn(tmp$eval, n)
plot(gk)
which.max(gk)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
tmp$dim
pairs(pred[,1:4],col=y+1)
# 2ndRKHS: gauss
plot(tmp$cv.shx)
max(tmp$cv.shx)
tmp$shx
# 2ndRKHS: poly
plot(tmp$cv.cd[-(139:140)])
max(tmp$cv.cd)
tmp$c;tmp$d
##########################
# Classification
##########################
set.seed(0)
train = sample(1:n, 100)
c.methods=c("lda", "qda", "svm")
d = tmp$dim
d=1
imethod=3
pred.class = pred
#pred.class = fpc.pred
out = classify(x=pred.class[train,1:d], y=y, x.test=pred.class[-train,1:d], method=c.methods[imethod])
mean(y[-train] != out)
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
pred.class = fpc.pred
out = classify(x=pred.class[train,1:d], y=y, x.test=pred.class[-train,1:d], method=c.methods[imethod])
mean(y[-train] != out)
# Linear FPCA (my code)
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear/FPCA")
#####################
# Visualization
#####################
out=data.frame(PC1=pred[,1], PC2 = pred[,2], Y=as.factor(y))
pdf('nfpc2d.pdf')
ggplot(out,aes(x=PC1,y=PC2, colour=Y)) +
geom_point()+theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('NAFPC'))
dev.off()
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
out=data.frame(PC1=fpc.pred[,1], PC2 = fpc.pred[,2], Y=as.factor(y))
pdf('fpc2d.pdf')
ggplot(out,aes(x=PC1,y=PC2, colour=Y)) +
geom_point()+theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('FPC'))
dev.off()
# curves - visualization of PC Scores
curve_df <- data.frame(id=1, time=tt, x = x[1,], nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x = x[i,], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
require(ggplot2)
p <- ggplot(curve_df, aes(x=time, y=x,group=id,colour=y))
pp <- p + geom_line() + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('Model I-1'), y='X(t)')
pdf('mi_description.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=time, y=x,group=id))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Model I (Original)'), y= 'X(t)')
#pdf('m1.pdf')
pp
#dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
# pdf('nfpc1.pdf')
pp
# dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc2,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd NAFPC'), y= 'X(t)')
# pdf('nfpc2.pdf')
pp
# dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('1st FPC'), y= 'X(t)')
pdf('fpc1.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc2,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd FPC'), y= 'X(t)')
pdf('fpc2.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), scaling(nfpc3,1)))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1), scaling(fpc3,1)))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), 0))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc2,1), 0,0))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), scaling(nfpc3,1)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 NAFPC'))
pdf('nfpc123.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1), scaling(fpc3,1)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 FPC'))
pdf('fpc123.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=time, y=x,group=id,colour=fpc1))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('1st (linear) FPC'))
pp
p <- ggplot(curve_df, aes(x=time, y=x,group=id,colour=fpc2))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd (linear) FPC'))
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1), scaling(fpc2), scaling(fpc3)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
#######################################################
# Model I-2 true function: linear -- just sum of a few eigenfunctions for each group
#######################################################
set.seed(0)
nt=20
n=200
nbasis=21
dat = sim.model.12(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
# Original Data
plot(tt,x[1,], type='l', ylim=c(-5,5), xlab="t", ylab="X(t)", main = "Model I", col=y)
for(i in 1:n)lines(tt,x[i,], col=(y[i]))
# Nonlinear FPCA
nafpca.m2=nafpca(x,tt, basisname="bspline", ex=0,nbasis=nbasis, gamma.tune=TRUE,ncv1=50,ncv2=50) # shx works
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='bspline')
nafpca.m2=nafpca(x,tt, nbasis=30,gamma.tune = TRUE, ncv1=30,ncv2=30,
basisname='fourier')
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=30,
basisname='gauss')
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=20,
basisname='brown')
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=20,
basisname='bspline', kernel='poly',
c=.1,d=2)
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='fourier', kernel='poly',
c=.5,d=2)
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='gauss', kernel='poly',
c=.5,d=2)
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='brown', kernel='poly',
c=.5,d=2)
pred=nafpca.m2$pred
pairs(pred[,1:4], col=y)
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
ncv2=length(nafpca.m2$cv.shx)
shx.grid = c(exp(seq(log(10^(-10)),log(10^(2)),len=ncv2)))
plot(shx.grid, nafpca.m2$cv.shx)
plot(Gn(nafpca.m2$eval,n))
which.max(Gn(nafpca.m2$eval,n))
cumsum(nafpca.m2$eval)/sum(nafpca.m2$eval)
nafpca.m2$shx
nafpca.m2$d
nafpca.m2$cv.cd
# Linear FPCA (my code)
x.ftn = get.fd(xraw=t(x),tt=tt,basisname='bspline',nbasis=nbasis,ncv=10)
fpc = fpca(x.ftn)
fpc.pred = fpc$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear/FPCA")
#####################
# Visualization
#####################
#y=y-1
out=data.frame(PC1=pred[,1], PC2 = pred[,2], Y=as.factor(y))
pdf('mi2_nfpc2d.pdf')
ggplot(out,aes(x=PC1,y=PC2, colour=Y)) +
geom_point()+theme_bw() +
scale_colour_manual(name="Y",
values = c("1"="black", "2"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('NAFPCA'))
dev.off()
# Linear FPCA
out=data.frame(PC1=fpc.pred[,1], PC2 = fpc.pred[,2], Y=as.factor(y))
pdf('mi2_fpc2d.pdf')
ggplot(out,aes(x=PC1,y=PC2, colour=Y)) +
geom_point()+theme_bw() +
scale_colour_manual(name="Y",
values = c("1"="black", "2"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('FPC'))
dev.off()
# curves - visualization of PC Scores
curve_df <- data.frame(id=1, time=tt, x = x[1,], nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x = x[i,], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
require(ggplot2)
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
p <- ggplot(curve_df, aes(x=time, y=x,group=id,colour=y))
pp <- p + geom_line() + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('Model I-2 with Y'), y='X(t)')
pdf('mi2_description.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=time, y=x,group=id))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Model I (Original)'), y= 'X(t)')
#pdf('mi2.pdf')
pp
#dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pdf('nfpc1.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc2,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd NAFPC'), y= 'X(t)')
pdf('nfpc2.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('1st FPC'), y= 'X(t)')
pdf('fpc1.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc2,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd FPC'), y= 'X(t)')
pdf('fpc2.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), scaling(nfpc3,1)))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1), scaling(fpc3,1)))
# hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), 0))
# hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc2,1), 0,0))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), scaling(nfpc3,1)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 NAFPC'))
pdf('nfpc123.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1), scaling(fpc3,1)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 FPC'))
pdf('fpc123.pdf')
pp
dev.off()
######################
# Model II-1 (unbalanced)
######################
set.seed(1)
n=200
nt=20
dat = sim.model.21(n=n, nt=nt, balanced = FALSE)
x = dat$x; tt = dat$tt; y = dat$y
i=which(y==0)[1]
plot(x[[1]][[i]], x[[2]][[i]], xlim=c(-5,5),ylim=c(-5,5), pch=16)
for(i in which(y==0)[1:10]){
points(x[[1]][[i]], x[[2]][[i]], xlim=c(-5,5),ylim=c(-5,5), pch=16)
}
for(j in which(y==1)[1:10]){
points(x[[1]][[j]], x[[2]][[j]], col=2, pch=16)
}
# Nonlinear FPCA
tmp=nafpca(x,tt, p=2, nbasis=21,unbalanced=TRUE, ncv1=10, ncv2=10)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
plot(tmp$cv.shx)
tmp$shx
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
# Linear FPCA (MFPCA/ PACE)
out = convert.fd(fd=x,tt=tt, p=2,n=n)
mfpc = MFPCA(out, M = 5, uniExpansions = list(list(type = "splines1D", k = 10),
list(type = "splines1D", k = 10)))
fpc.pred = mfpc$scores
plotpred(fpc.pred, vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear/PACE")
curve_df <- data.frame(id=1, time=tt, x1 = x[[1]][[1]], x2=x[[2]][[1]],nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[[1]][[i]], x2=x[[2]][[i]], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
require(ggplot2)
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=y))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=nfpc1))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=fpc1))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
######################
# Model II-1 (balanced)
######################
n=100
nt=10
dat = sim.model.21(n=n, nt=nt, balanced = TRUE)
x = dat$x; tt = dat$tt; y = dat$y
y
i=which(y==0)[1]
plot(x[i,,1], x[i,,2], col=1, xlim=c(-5,5),ylim=c(-5,5), pch=16)
for(i in which(y==0)[1:10]){
points(x[i,,1], x[i,,2], col=1, xlim=c(-5,5),ylim=c(-5,5), pch=16)
}
for(j in which(y==1)[1:10]){
points(x[j,,1], x[j,,2], col=2, pch=16)
}
# Nonlinear FPCA
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='bspline')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='fourier')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='brown')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='gauss')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100, kernel = 'poly', d=1,
basisname='bspline')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100, kernel = 'poly', d=3,
basisname='fourier')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100, kernel = 'poly', d=3,
basisname='brown')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100, kernel = 'poly', d=3,
basisname='gauss')
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
#plot(tmp$cv.shx)
max(tmp$cv.shx)
gk = Gn(tmp$eval, n)
#plot(gk)
which.max(gk)
plot(tmp$cv.cd)
max(tmp$cv.cd)
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
# Linear FPCA (MFPCA/ PACE)
# out = convert.fd(fd=x,tt=tt, p=2,n=n)
# mfpc = MFPCA(out, M = 5, uniExpansions = list(list(type = "splines1D", k = 10),
# list(type = "splines1D", k = 10)))
# fpc.pred = mfpc$scores
#
# plotpred(fpc.pred, vec=vec,ind.inf=y,
# xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
# main="Linear/PACE")
curve_df <- data.frame(id=1, time=tt, x1 = x[[1]][[1]], x2=x[[2]][[1]],nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[[1]][[i]], x2=x[[2]][[i]], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
require(ggplot2)
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=y))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=nfpc1))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=fpc1))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
######################
# Model II-2 (unbalanced)
######################
n=200
nt=20
set.seed(0)
dat = sim.model.22(n=n, nt=nt, p=2, balanced=FALSE, m1=1, m2=2, r1=1, r2=0.7, sd1=.4, sd2=.2)
x = dat$x; tt = dat$tt; y = dat$y
i=which(y==0)[1]
plot(x[[1]][[i]], x[[2]][[i]], xlim=c(-5,5),ylim=c(-5,5), pch=16)
for(i in which(y==0)[1:10]){
points(x[[1]][[i]], x[[2]][[i]], xlim=c(-5,5),ylim=c(-5,5), pch=16)
}
for(j in which(y==1)[1:10]){
points(x[[1]][[j]], x[[2]][[j]], col=2, pch=16)
}
# Nonlinear FPCA
tmp=nafpca(x,tt, p=2, nbasis=8,unbalanced=TRUE,ncv1=20, ncv2=20, basisname='bspline', gamma.tune=TRUE,)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
plot(tmp$cv.shx)
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
# Linear FPCA (MFPCA/ PACE)
# out = convert.fd(fd=x,tt=tt, p=2,n=n)
# mfpc = MFPCA(out, M = 5, uniExpansions = list(list(type = "splines1D", k = 10),
# list(type = "splines1D", k = 10)))
# fpc.pred = mfpc$scores
#
# plotpred(fpc.pred, vec=vec,ind.inf=y,
# xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
# main="Linear/PACE")
curve_df <- data.frame(id=1, time=tt, x1 = x[[1]][[1]], x2=x[[2]][[1]],nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[[1]][[i]], x2=x[[2]][[i]], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
curve_df$ry = rgb(scaling(as.numeric(curve_df$y),1),0,0)
require(ggplot2)
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=y))
pp <- p + geom_point(aes(colour=curve_df$y)) + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with Y'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pdf('m2.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id))
pp <- p + geom_point() + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II (Original)'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pdf('m2_original.pdf')
pp
dev.off()
out = data.frame(PC1=pred[,1], PC2=pred[,2], Y=as.factor(y))
p <- ggplot(out, aes(x=PC1, y=PC2, colour=Y))
pp <- p + geom_point(aes(colour=Y)) + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('NAFPC'))
pdf('m2nfpc2d.pdf')
pp
dev.off()
out = data.frame(PC1=fpc.pred[,1], PC2=fpc.pred[,2], Y=as.factor(y))
p <- ggplot(out, aes(x=PC1, y=PC2, colour=Y))
pp <- p + geom_point(aes(colour=Y)) + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('FPC'))
pdf('m2fpc2d.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=nfpc1))
pp <- p + geom_point(aes(colour=nfpc1)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pdf('m2nfpc1.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=fpc1))
pp <- p + geom_point(aes(colour=fpc1)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with FPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pdf('m2fpc1.pdf')
pp
dev.off()
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1),0,0))
p <- ggplot(hex.df, aes(x=x1, y=x2,group=id,colour=hex))
pp <- p + geom_point(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 linear FPC'))
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0, 0))
p <- ggplot(hex.df, aes(x=x1, y=x2,group=id,colour=hex))
pp <- p + geom_point(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 linear FPC'))
pp
#############################################
# Model I-3 : classification
#######################################################
set.seed(0)
nt=20
n=200
nbasis=21
dat = sim.model.31(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
# Original Data
plot(tt,x[1,], type='l', ylim=c(min(x),max(x)), xlab="t", ylab="X(t)", main = "Model I", col=y)
for(i in 1:n)lines(tt,x[i,], col=(y[i]))
# Nonlinear FPCA
tmp=nafpca(x,tt, basisname="bspline",nbasis=nbasis, gamma.tune=TRUE,ncv1=20,ncv2=20) # shx works
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
tmp$dim
aa=fpca(nafpca.m2$ftn)
pairs(pred[,1:10],col=y)
pairs(aa$pred[,1:10],col=y)
######################
# Model II-3
######################
set.seed(0)
n=200
nt=20
nbasis=21
dat = sim.model.23(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
fdplot(x,y,tt, n.point=100)
i=which(y==1)[1]
plot(x[i,,1], x[i,,2], col=1, xlim=c(-5,5),ylim=c(-5,5), pch=16)
for(i in which(y==1)[1:10]){
points(x[i,,1], x[i,,2], col=1, xlim=c(-5,5),ylim=c(-5,5), pch=16)
}
for(j in which(y==2)[1:10]){
points(x[j,,1], x[j,,2], col=2, pch=16)
}
# Nonlinear FPCA
tmp=nafpca(x,tt, p=2, nbasis=nbasis, ncv1= 30, ncv2=30)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
pairs(pred[,1:4], col=y)
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
######################
# Model III-2
######################
set.seed(0)
n=200
nt=20
nbasis=21
dat = sim.model.32(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
fdplot(x,y,tt)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=1, nbasis=nbasis, ncv1= 50, ncv2=50)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
pairs(pred[,1:4], col=y)
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
######################
# Model III-3
######################
set.seed(0)
n=200
nt=20
nbasis=21
dat = sim.model.33(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
fdplot(x,y,tt)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=1, nbasis=nbasis, ncv1= 50, ncv2=50)
pred=tmp$pred
vec=c(3,4)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
pairs(pred[,1:6], col=y)
#pairs(pred[,1:4], col=y)
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
######################
# Model III-4
######################
set.seed(0)
n=200
nt=20
nbasis=31
dat = sim.model.34(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
fdplot(x,y,tt)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=1, nbasis=nbasis, ncv1= 50, ncv2=50)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
pairs(pred[,1:6], col=y)
#pairs(pred[,1:4], col=y)
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
######################
# Model III-5
######################
set.seed(0)
n=200
nt=20
nbasis=31
dat = sim.model.35(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
y=as.factor(y)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=3, nbasis=nbasis, ncv1= 50, ncv2=50)
pred=tmp$pred
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
summary(glm(y~pred[,1:5], family = binomial(link="logit")))
summary(glm(y~fpc.pred[,1:5],family = binomial(link="logit")))
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
i=1
curve_df <- data.frame(id=1, time=tt, x1 = x[1,,1], x2=x[1,,2], x3=x[1,,3],nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[i,,1], x2=x[i,,2], x3=x[i,,3],nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
p <- ggplot(curve_df, aes(x=nfpc1, y=nfpc2,group=id,colour=y))
pp <- p + geom_point(aes(colour=y)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pp
p <- ggplot(curve_df, aes(x=fpc1, y=fpc2,group=id,colour=y))
pp <- p + geom_point(aes(colour=y)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc3,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st FPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1),scaling(nfpc3,1),))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1),scaling(fpc3,1),))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
######################
# Model III-6
######################
set.seed(0)
n=200
nt=20
nbasis=21
dat = sim.model.36(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
y=as.factor(y)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=2, nbasis=nbasis, ncv1= 10, ncv2=10)
pred=tmp$pred
plot(x[6,,1])
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
summary(glm(y~pred[,1:5], family = binomial(link="logit")))
summary(glm(y~fpc.pred[,1:5],family = binomial(link="logit")))
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
i=1
curve_df <- data.frame(id=1, time=tt, x1 = x[1,,1], x2=x[1,,2], nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[i,,1], x2=x[i,,2],nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=nfpc3))
pp <- p + geom_point(aes(colour=nfpc3)) + theme_bw() +#scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pp
p <- ggplot(curve_df, aes(x=fpc1, y=fpc2,group=id,colour=y))
pp <- p + geom_point(aes(colour=y)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc3,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st FPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1),scaling(nfpc3,1),))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1),scaling(fpc3,1),))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
| /nafpca/simulation/sim_plots.R | no_license | CodeJSong/Functional-Dimension-Reduction | R | false | false | 37,420 | r | source('simulation/sim_model.R')
#install.packages(c("MFPCA","funData"))
#install.packages("funData")
source('nafpca.R')
source('../../classification.R')
library(MFPCA);library(funData)
library(dplyr)
library(ggplot2)
scaling <- function(x,k) {
#x = abs(x)
out = (x - min(x)) / (max(x) - min(x)) / k
return(out)
}
##################################################################
# Model I-1 : 1-d functional data - nonlinear relation in clustering
##################################################################
n=200; nt=20
set.seed(1)
dat = sim.model.11(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
# Original Data
plot(tt,x[1,], type='l', ylim=c(-5,5), xlab="t", ylab="X(t)", main = "Model I", col=y+1)
for(i in 1:n)lines(tt,x[i,], col=(y[i]+1))
# function estimation
# 1. fda-based
# bspline: most natural
coef.bspline = get.fd(t(x),tt ,basisname='bspline')
fd.bsp = fd(coef.bspline$coef, coef.bspline$basis)
plot(fd.bsp, col=y+1)
evals = t(eval.fd(tt,fd.bsp))
sum((evals - x)^2) / n /nt
# fourier: tend to get x[0] = 0 , x[1] =0
coef.fourier = get.fd(t(x),tt ,basisname='fourier', nbasis=53)
fd.fourier=fd(coef.fourier$coef, coef.fourier$basis)
plot(fd.fourier, col=y+1)
evals = t(eval.fd(tt,fd.fourier))
sum((evals - x)^2) / n /nt
# 2. rkhs-based
# gaussian: natural
coef.gauss = get.fd.rkhs(x, tt, kern="gauss")
evals = coef.gauss$coef %*% coef.gauss$kt
plot(tt, evals[1,], col=y[1]+1, type='l', ylim=c(min(evals),max(evals)))
for(j in 1:n){
lines(tt, evals[j,], col=y[j]+1)
}
sum((evals - x)^2) / n /nt
# brownian: natural --- the best in this situation
coef.brown = get.fd.rkhs(x, tt, kern="brown")
evals = coef.brown$coef %*% coef.brown$kt
plot(tt, evals[1,], col=y[1]+1, type='l', ylim=c(min(evals),max(evals)))
for(j in 1:n){
lines(tt, evals[j,], col=y[j]+1)
}
sum((evals - x)^2) / n /nt
# Nonlinear FPCA
tmp=nafpca(x,tt, shx=11, nbasis=21,gamma.tune = TRUE,
basisname='bspline')
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='fourier')
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=30,
basisname='gauss')
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=20,
basisname='brown')
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=20,
basisname='bspline', kernel='poly',
c=.1,d=2)
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='fourier', kernel='poly',
c=.5,d=2)
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='gauss', kernel='poly',
c=.5,d=2)
tmp=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='brown', kernel='poly',
c=.5,d=2)
tmp$eval
gk = Gn(tmp$eval, n)
plot(gk)
which.max(gk)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
tmp$dim
pairs(pred[,1:4],col=y+1)
# 2ndRKHS: gauss
plot(tmp$cv.shx)
max(tmp$cv.shx)
tmp$shx
# 2ndRKHS: poly
plot(tmp$cv.cd[-(139:140)])
max(tmp$cv.cd)
tmp$c;tmp$d
##########################
# Classification
##########################
set.seed(0)
train = sample(1:n, 100)
c.methods=c("lda", "qda", "svm")
d = tmp$dim
d=1
imethod=3
pred.class = pred
#pred.class = fpc.pred
out = classify(x=pred.class[train,1:d], y=y, x.test=pred.class[-train,1:d], method=c.methods[imethod])
mean(y[-train] != out)
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
pred.class = fpc.pred
out = classify(x=pred.class[train,1:d], y=y, x.test=pred.class[-train,1:d], method=c.methods[imethod])
mean(y[-train] != out)
# Linear FPCA (my code)
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear/FPCA")
#####################
# Visualization
#####################
out=data.frame(PC1=pred[,1], PC2 = pred[,2], Y=as.factor(y))
pdf('nfpc2d.pdf')
ggplot(out,aes(x=PC1,y=PC2, colour=Y)) +
geom_point()+theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('NAFPC'))
dev.off()
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
out=data.frame(PC1=fpc.pred[,1], PC2 = fpc.pred[,2], Y=as.factor(y))
pdf('fpc2d.pdf')
ggplot(out,aes(x=PC1,y=PC2, colour=Y)) +
geom_point()+theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('FPC'))
dev.off()
# curves - visualization of PC Scores
curve_df <- data.frame(id=1, time=tt, x = x[1,], nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x = x[i,], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
require(ggplot2)
p <- ggplot(curve_df, aes(x=time, y=x,group=id,colour=y))
pp <- p + geom_line() + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('Model I-1'), y='X(t)')
pdf('mi_description.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=time, y=x,group=id))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Model I (Original)'), y= 'X(t)')
#pdf('m1.pdf')
pp
#dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
# pdf('nfpc1.pdf')
pp
# dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc2,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd NAFPC'), y= 'X(t)')
# pdf('nfpc2.pdf')
pp
# dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('1st FPC'), y= 'X(t)')
pdf('fpc1.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc2,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd FPC'), y= 'X(t)')
pdf('fpc2.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), scaling(nfpc3,1)))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1), scaling(fpc3,1)))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), 0))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc2,1), 0,0))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), scaling(nfpc3,1)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 NAFPC'))
pdf('nfpc123.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1), scaling(fpc3,1)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 FPC'))
pdf('fpc123.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=time, y=x,group=id,colour=fpc1))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('1st (linear) FPC'))
pp
p <- ggplot(curve_df, aes(x=time, y=x,group=id,colour=fpc2))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd (linear) FPC'))
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1), scaling(fpc2), scaling(fpc3)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
#######################################################
# Model I-2 true function: linear -- just sum of a few eigenfunctions for each group
#######################################################
set.seed(0)
nt=20
n=200
nbasis=21
dat = sim.model.12(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
# Original Data
plot(tt,x[1,], type='l', ylim=c(-5,5), xlab="t", ylab="X(t)", main = "Model I", col=y)
for(i in 1:n)lines(tt,x[i,], col=(y[i]))
# Nonlinear FPCA
nafpca.m2=nafpca(x,tt, basisname="bspline", ex=0,nbasis=nbasis, gamma.tune=TRUE,ncv1=50,ncv2=50) # shx works
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='bspline')
nafpca.m2=nafpca(x,tt, nbasis=30,gamma.tune = TRUE, ncv1=30,ncv2=30,
basisname='fourier')
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=30,
basisname='gauss')
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=20,
basisname='brown')
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,ncv1=30,ncv2=20,
basisname='bspline', kernel='poly',
c=.1,d=2)
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='fourier', kernel='poly',
c=.5,d=2)
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='gauss', kernel='poly',
c=.5,d=2)
nafpca.m2=nafpca(x,tt, shx=11, nbasis=30,gamma.tune = TRUE,
basisname='brown', kernel='poly',
c=.5,d=2)
pred=nafpca.m2$pred
pairs(pred[,1:4], col=y)
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
ncv2=length(nafpca.m2$cv.shx)
shx.grid = c(exp(seq(log(10^(-10)),log(10^(2)),len=ncv2)))
plot(shx.grid, nafpca.m2$cv.shx)
plot(Gn(nafpca.m2$eval,n))
which.max(Gn(nafpca.m2$eval,n))
cumsum(nafpca.m2$eval)/sum(nafpca.m2$eval)
nafpca.m2$shx
nafpca.m2$d
nafpca.m2$cv.cd
# Linear FPCA (my code)
x.ftn = get.fd(xraw=t(x),tt=tt,basisname='bspline',nbasis=nbasis,ncv=10)
fpc = fpca(x.ftn)
fpc.pred = fpc$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear/FPCA")
#####################
# Visualization
#####################
#y=y-1
out=data.frame(PC1=pred[,1], PC2 = pred[,2], Y=as.factor(y))
pdf('mi2_nfpc2d.pdf')
ggplot(out,aes(x=PC1,y=PC2, colour=Y)) +
geom_point()+theme_bw() +
scale_colour_manual(name="Y",
values = c("1"="black", "2"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('NAFPCA'))
dev.off()
# Linear FPCA
out=data.frame(PC1=fpc.pred[,1], PC2 = fpc.pred[,2], Y=as.factor(y))
pdf('mi2_fpc2d.pdf')
ggplot(out,aes(x=PC1,y=PC2, colour=Y)) +
geom_point()+theme_bw() +
scale_colour_manual(name="Y",
values = c("1"="black", "2"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('FPC'))
dev.off()
# curves - visualization of PC Scores
curve_df <- data.frame(id=1, time=tt, x = x[1,], nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x = x[i,], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
require(ggplot2)
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
p <- ggplot(curve_df, aes(x=time, y=x,group=id,colour=y))
pp <- p + geom_line() + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=20),
legend.position="right")+
theme(plot.title = element_text(hjust = 0.5,size=20))+
labs(title = paste0('Model I-2 with Y'), y='X(t)')
pdf('mi2_description.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=time, y=x,group=id))
pp <- p + geom_line() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Model I (Original)'), y= 'X(t)')
#pdf('mi2.pdf')
pp
#dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pdf('nfpc1.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc2,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd NAFPC'), y= 'X(t)')
pdf('nfpc2.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('1st FPC'), y= 'X(t)')
pdf('fpc1.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc2,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('2nd FPC'), y= 'X(t)')
pdf('fpc2.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), scaling(nfpc3,1)))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1), scaling(fpc3,1)))
# hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), 0))
# hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc2,1), 0,0))
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1), scaling(nfpc3,1)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 NAFPC'))
pdf('nfpc123.pdf')
pp
dev.off()
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1), scaling(fpc3,1)))
p <- ggplot(hex.df, aes(x=time, y=x,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 FPC'))
pdf('fpc123.pdf')
pp
dev.off()
######################
# Model II-1 (unbalanced)
######################
set.seed(1)
n=200
nt=20
dat = sim.model.21(n=n, nt=nt, balanced = FALSE)
x = dat$x; tt = dat$tt; y = dat$y
i=which(y==0)[1]
plot(x[[1]][[i]], x[[2]][[i]], xlim=c(-5,5),ylim=c(-5,5), pch=16)
for(i in which(y==0)[1:10]){
points(x[[1]][[i]], x[[2]][[i]], xlim=c(-5,5),ylim=c(-5,5), pch=16)
}
for(j in which(y==1)[1:10]){
points(x[[1]][[j]], x[[2]][[j]], col=2, pch=16)
}
# Nonlinear FPCA
tmp=nafpca(x,tt, p=2, nbasis=21,unbalanced=TRUE, ncv1=10, ncv2=10)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
plot(tmp$cv.shx)
tmp$shx
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
# Linear FPCA (MFPCA/ PACE)
out = convert.fd(fd=x,tt=tt, p=2,n=n)
mfpc = MFPCA(out, M = 5, uniExpansions = list(list(type = "splines1D", k = 10),
list(type = "splines1D", k = 10)))
fpc.pred = mfpc$scores
plotpred(fpc.pred, vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear/PACE")
curve_df <- data.frame(id=1, time=tt, x1 = x[[1]][[1]], x2=x[[2]][[1]],nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[[1]][[i]], x2=x[[2]][[i]], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
require(ggplot2)
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=y))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=nfpc1))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=fpc1))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
######################
# Model II-1 (balanced)
######################
n=100
nt=10
dat = sim.model.21(n=n, nt=nt, balanced = TRUE)
x = dat$x; tt = dat$tt; y = dat$y
y
i=which(y==0)[1]
plot(x[i,,1], x[i,,2], col=1, xlim=c(-5,5),ylim=c(-5,5), pch=16)
for(i in which(y==0)[1:10]){
points(x[i,,1], x[i,,2], col=1, xlim=c(-5,5),ylim=c(-5,5), pch=16)
}
for(j in which(y==1)[1:10]){
points(x[j,,1], x[j,,2], col=2, pch=16)
}
# Nonlinear FPCA
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='bspline')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='fourier')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='brown')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100,
basisname='gauss')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100, kernel = 'poly', d=1,
basisname='bspline')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100, kernel = 'poly', d=3,
basisname='fourier')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100, kernel = 'poly', d=3,
basisname='brown')
tmp=nafpca(x,tt,nbasis=30,gamma.tune = TRUE, ncv2=100, kernel = 'poly', d=3,
basisname='gauss')
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
#plot(tmp$cv.shx)
max(tmp$cv.shx)
gk = Gn(tmp$eval, n)
#plot(gk)
which.max(gk)
plot(tmp$cv.cd)
max(tmp$cv.cd)
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
# Linear FPCA (MFPCA/ PACE)
# out = convert.fd(fd=x,tt=tt, p=2,n=n)
# mfpc = MFPCA(out, M = 5, uniExpansions = list(list(type = "splines1D", k = 10),
# list(type = "splines1D", k = 10)))
# fpc.pred = mfpc$scores
#
# plotpred(fpc.pred, vec=vec,ind.inf=y,
# xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
# main="Linear/PACE")
curve_df <- data.frame(id=1, time=tt, x1 = x[[1]][[1]], x2=x[[2]][[1]],nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[[1]][[i]], x2=x[[2]][[i]], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
require(ggplot2)
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=y))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=nfpc1))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=fpc1))
pp <- p + geom_point() + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('Sparsely observed'))
pp
######################
# Model II-2 (unbalanced)
######################
n=200
nt=20
set.seed(0)
dat = sim.model.22(n=n, nt=nt, p=2, balanced=FALSE, m1=1, m2=2, r1=1, r2=0.7, sd1=.4, sd2=.2)
x = dat$x; tt = dat$tt; y = dat$y
i=which(y==0)[1]
plot(x[[1]][[i]], x[[2]][[i]], xlim=c(-5,5),ylim=c(-5,5), pch=16)
for(i in which(y==0)[1:10]){
points(x[[1]][[i]], x[[2]][[i]], xlim=c(-5,5),ylim=c(-5,5), pch=16)
}
for(j in which(y==1)[1:10]){
points(x[[1]][[j]], x[[2]][[j]], col=2, pch=16)
}
# Nonlinear FPCA
tmp=nafpca(x,tt, p=2, nbasis=8,unbalanced=TRUE,ncv1=20, ncv2=20, basisname='bspline', gamma.tune=TRUE,)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
plot(tmp$cv.shx)
# Linear FPCA
fpc = fpca(tmp$ftn)
fpc.pred = fpc$pred
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
# Linear FPCA (MFPCA/ PACE)
# out = convert.fd(fd=x,tt=tt, p=2,n=n)
# mfpc = MFPCA(out, M = 5, uniExpansions = list(list(type = "splines1D", k = 10),
# list(type = "splines1D", k = 10)))
# fpc.pred = mfpc$scores
#
# plotpred(fpc.pred, vec=vec,ind.inf=y,
# xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
# main="Linear/PACE")
curve_df <- data.frame(id=1, time=tt, x1 = x[[1]][[1]], x2=x[[2]][[1]],nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[[1]][[i]], x2=x[[2]][[i]], nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
curve_df$id <- factor(curve_df$id)
curve_df$y <- factor(curve_df$y)
curve_df$ry = rgb(scaling(as.numeric(curve_df$y),1),0,0)
require(ggplot2)
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=y))
pp <- p + geom_point(aes(colour=curve_df$y)) + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with Y'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pdf('m2.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id))
pp <- p + geom_point() + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II (Original)'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pdf('m2_original.pdf')
pp
dev.off()
out = data.frame(PC1=pred[,1], PC2=pred[,2], Y=as.factor(y))
p <- ggplot(out, aes(x=PC1, y=PC2, colour=Y))
pp <- p + geom_point(aes(colour=Y)) + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('NAFPC'))
pdf('m2nfpc2d.pdf')
pp
dev.off()
out = data.frame(PC1=fpc.pred[,1], PC2=fpc.pred[,2], Y=as.factor(y))
p <- ggplot(out, aes(x=PC1, y=PC2, colour=Y))
pp <- p + geom_point(aes(colour=Y)) + theme_bw() +
scale_colour_manual(name="Y",
values = c("0"="black", "1"="red"))+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('FPC'))
pdf('m2fpc2d.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=nfpc1))
pp <- p + geom_point(aes(colour=nfpc1)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pdf('m2nfpc1.pdf')
pp
dev.off()
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=fpc1))
pp <- p + geom_point(aes(colour=fpc1)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with FPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pdf('m2fpc1.pdf')
pp
dev.off()
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1),0,0))
p <- ggplot(hex.df, aes(x=x1, y=x2,group=id,colour=hex))
pp <- p + geom_point(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 linear FPC'))
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0, 0))
p <- ggplot(hex.df, aes(x=x1, y=x2,group=id,colour=hex))
pp <- p + geom_point(colour=hex.df$hex) + theme_bw() +
theme(legend.position="none") +
theme(plot.title = element_text(hjust = 0.5,size=22))+
labs(title = paste0('First 3 linear FPC'))
pp
#############################################
# Model I-3 : classification
#######################################################
set.seed(0)
nt=20
n=200
nbasis=21
dat = sim.model.31(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
# Original Data
plot(tt,x[1,], type='l', ylim=c(min(x),max(x)), xlab="t", ylab="X(t)", main = "Model I", col=y)
for(i in 1:n)lines(tt,x[i,], col=(y[i]))
# Nonlinear FPCA
tmp=nafpca(x,tt, basisname="bspline",nbasis=nbasis, gamma.tune=TRUE,ncv1=20,ncv2=20) # shx works
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
tmp$dim
aa=fpca(nafpca.m2$ftn)
pairs(pred[,1:10],col=y)
pairs(aa$pred[,1:10],col=y)
######################
# Model II-3
######################
set.seed(0)
n=200
nt=20
nbasis=21
dat = sim.model.23(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
fdplot(x,y,tt, n.point=100)
i=which(y==1)[1]
plot(x[i,,1], x[i,,2], col=1, xlim=c(-5,5),ylim=c(-5,5), pch=16)
for(i in which(y==1)[1:10]){
points(x[i,,1], x[i,,2], col=1, xlim=c(-5,5),ylim=c(-5,5), pch=16)
}
for(j in which(y==2)[1:10]){
points(x[j,,1], x[j,,2], col=2, pch=16)
}
# Nonlinear FPCA
tmp=nafpca(x,tt, p=2, nbasis=nbasis, ncv1= 30, ncv2=30)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
pairs(pred[,1:4], col=y)
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
######################
# Model III-2
######################
set.seed(0)
n=200
nt=20
nbasis=21
dat = sim.model.32(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
fdplot(x,y,tt)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=1, nbasis=nbasis, ncv1= 50, ncv2=50)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
pairs(pred[,1:4], col=y)
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
######################
# Model III-3
######################
set.seed(0)
n=200
nt=20
nbasis=21
dat = sim.model.33(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
fdplot(x,y,tt)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=1, nbasis=nbasis, ncv1= 50, ncv2=50)
pred=tmp$pred
vec=c(3,4)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
pairs(pred[,1:6], col=y)
#pairs(pred[,1:4], col=y)
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
######################
# Model III-4
######################
set.seed(0)
n=200
nt=20
nbasis=31
dat = sim.model.34(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
fdplot(x,y,tt)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=1, nbasis=nbasis, ncv1= 50, ncv2=50)
pred=tmp$pred
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
pairs(pred[,1:6], col=y)
#pairs(pred[,1:4], col=y)
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
vec=c(1,2)
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
######################
# Model III-5
######################
set.seed(0)
n=200
nt=20
nbasis=31
dat = sim.model.35(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
y=as.factor(y)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=3, nbasis=nbasis, ncv1= 50, ncv2=50)
pred=tmp$pred
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
summary(glm(y~pred[,1:5], family = binomial(link="logit")))
summary(glm(y~fpc.pred[,1:5],family = binomial(link="logit")))
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
i=1
curve_df <- data.frame(id=1, time=tt, x1 = x[1,,1], x2=x[1,,2], x3=x[1,,3],nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[i,,1], x2=x[i,,2], x3=x[i,,3],nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
p <- ggplot(curve_df, aes(x=nfpc1, y=nfpc2,group=id,colour=y))
pp <- p + geom_point(aes(colour=y)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pp
p <- ggplot(curve_df, aes(x=fpc1, y=fpc2,group=id,colour=y))
pp <- p + geom_point(aes(colour=y)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc3,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st FPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1),scaling(nfpc3,1),))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1),scaling(fpc3,1),))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
######################
# Model III-6
######################
set.seed(0)
n=200
nt=20
nbasis=21
dat = sim.model.36(n=n, nt=nt)
x = dat$x; tt = dat$tt; y = dat$y
y=as.factor(y)
# Nonlinear FPCA
tmp=nafpca(x,tt, p=2, nbasis=nbasis, ncv1= 10, ncv2=10)
pred=tmp$pred
plot(x[6,,1])
fpca.out = fpca(tmp$ftn)
fpc.pred = fpca.out$pred
summary(glm(y~pred[,1:5], family = binomial(link="logit")))
summary(glm(y~fpc.pred[,1:5],family = binomial(link="logit")))
vec=c(1,2)
plotpred(pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Nonlinear")
plotpred(fpc.pred,vec=vec,ind.inf=y,
xlab=paste("PC ",vec[1]), ylab=paste("PC ",vec[2]),
main="Linear")
i=1
curve_df <- data.frame(id=1, time=tt, x1 = x[1,,1], x2=x[1,,2], nfpc1=pred[1,1], nfpc2=pred[1,2], nfpc3=pred[1,3],
fpc1=fpc.pred[1,1], fpc2=fpc.pred[1,2], fpc3=fpc.pred[1,3], y=y[i])
for(i in 2:n){
curve_df <- rbind.data.frame(curve_df, data.frame(id=i, time=tt, x1 = x[i,,1], x2=x[i,,2],nfpc1=pred[i,1], nfpc2=pred[i,2], nfpc3=pred[i,3],
fpc1=fpc.pred[i,1], fpc2=fpc.pred[i,2], fpc3=fpc.pred[i,3], y=y[i]))
}
p <- ggplot(curve_df, aes(x=x1, y=x2,group=id,colour=nfpc3))
pp <- p + geom_point(aes(colour=nfpc3)) + theme_bw() +#scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pp
p <- ggplot(curve_df, aes(x=fpc1, y=fpc2,group=id,colour=y))
pp <- p + geom_point(aes(colour=y)) + theme_bw() +scale_colour_gradient(low="black",high="red")+
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="right")+
labs(title = paste0('Model II with NAFPC'),
x = expression({X^{1}}(t)),
y = expression({X^{2}}(t)))
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc3,1), 0,0))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st FPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(nfpc1,1), scaling(nfpc2,1),scaling(nfpc3,1),))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
hex.df <- curve_df %>% mutate(hex = rgb(scaling(fpc1,1), scaling(fpc2,1),scaling(fpc3,1),))
p <- ggplot(hex.df, aes(x=time, y=x3,group=id,colour=hex))
pp <- p + geom_line(colour=hex.df$hex) + theme_bw() +
theme(plot.title = element_text(hjust = 0.5,size=22),
legend.position="none") +
labs(title = paste0('1st NAFPC'), y= 'X(t)')
pp
|
require(readr)
require(plyr)
require(igraph)
require(rgexf)
# set working directory
getwd()
setwd("../query_results/merge_scripts/complete_merge/")
# read node and edges into dataframe with the name expected by igraph
nodes <- read.csv("time_slice_1536_complete_merge_budΓ©_and_era_correspondents.csv", fileEncoding="UTF-8")
links <- read.csv("time_slice_1536_complete_merge_budΓ©_and_era_letters_corr_as_nodes.csv", fileEncoding="UTF-8")[ ,c('Source', 'Target')]
setwd("../../")
getwd()
mutcorr <- read.csv("./intersection_overview/id_and_names_of_mut_corr_era_budΓ©.csv", fileEncoding="UTF-8")
# add colour for all correspondents
nodes$colour <- "#525252"
# add colour column for mutual correspondents t
nodes$colour <- ifelse(nodes$Id %in% mutcorr$correspondents_id, as.character("#C3161F"), nodes$colour)
#assign specific colour for erasmus
nodes$colour <- ifelse(nodes$Id == "17c580aa-3ba7-4851-8f26-9b3a0ebeadbf", as.character("#3C93AF"), nodes$colour)
#assign specific colour for budΓ©
nodes$colour <- ifelse(nodes$Id == "c0b89c75-45b8-4b04-bfd7-25bfe9ed040b", as.character("#D5AB5B"), nodes$colour)
#assign edge weight
links$weight <- 1
# create igraph object
net <- graph_from_data_frame(d=links, vertices=nodes, directed=T)
# conduct edge bundling (sum edge weights)
net2 <- simplify(net, remove.multiple = TRUE, edge.attr.comb=list(weight="sum","ignore"))
# calculate degree for all nodes
degAll <- degree(net2, v = V(net2), mode = "all")
# calculate weighted degree for all nodes
weightDegAll <- strength(net2, vids = V(net2), mode = "all",
loops = TRUE)
# add new node and edge attributes based on the calculated properties, add
net2 <- set.vertex.attribute(net2, "weightDegAll", index = V(net2), value = weightDegAll)
net2 <- set.vertex.attribute(net2, "degree", index = V(net2), value = degAll)
net2 <- set.vertex.attribute(net2, "colour", index = V(net2), value = nodes$colour)
net2 <- set.edge.attribute(net2, "weight", index = E(net2), value = E(net2)$weight)
#assign edge colour according to source node
edge.start <- ends(net2, es=E(net2), names=F)[,1]
edge.col <- V(net2)$colour[edge.start]
# layout with FR
l <- layout_with_fr(net2, weights=E(net2)$weight)*3.5
# plot graph
plot(net2, layout=l*5, vertex.color=nodes$colour, vertex.size=2, vertex.label=V(net2)$Label, vertex.label.font=2, vertex.label.color="gray40",
vertex.label.cex=.3, edge.arrow.size=.2, edge.width=E(net2)$weight*0.5, edge.color=edge.col, vertex.label.family="sans")
#################
# calculate node coordinates
nodes_coord <- as.data.frame(layout.fruchterman.reingold(net2, weights=E(net2)$weight)*50)
nodes_coord <- cbind(nodes_coord, rep(0, times = nrow(nodes_coord)))
# assign a colour for each node
nodes_col <- V(net2)$colour
# transform nodes into a data frame
nodes_col_df <- as.data.frame(t(col2rgb(nodes_col, alpha = FALSE)))
nodes_col_df <- cbind(nodes_col_df, alpha = rep(1, times = nrow(nodes_col_df)))
# assign visual attributes to nodes (RGBA)
nodes_att_viz <- list(color = nodes_col_df, position = nodes_coord)
# assign a colour for each edge
edges_col <- edge.col
# Transform it into a data frame (we have to transpose it first)
edges_col_df <- as.data.frame(t(col2rgb(edges_col, alpha = FALSE)))
edges_col_df <- cbind(edges_col_df, alpha = rep(1, times = nrow(edges_col_df)))
# assign visual attributes to edges (RGBA)
edges_att_viz <- list(color = edges_col_df)
# create data frames for gexf export
nodes_df <- data.frame(ID = c(1:vcount(net2)), NAME = V(net2)$Label)
edges_df <- as.data.frame(get.edges(net2, c(1:ecount(net2))))
#create a dataframe with node attributes
nodes_att <- data.frame(Degree = V(net2)$degree, colour = as.character(nodes$colour), "Weighted Degree" = V(net2)$weightDegAll)
setwd("../")
getwd()
setwd("./network_data/complete_merge_time_slices_gexf_created_by_r")
# write gexf
era_budΓ©_cmerge_1536 <- write.gexf(nodes = nodes_df, edges = edges_df, edgesWeight = E(net2)$weight, nodesAtt = nodes_att, nodesVizAtt = nodes_att_viz, edgesVizAtt = edges_att_viz, defaultedgetype = "directed", meta = list( creator="Christoph Kudella", description="A graph representing the intersection between Erasmus's and BudΓ©'s networks of correspondence in the year 1536"), output="era_budΓ©_cmerge_1536.gexf") | /intersections_budΓ©_erasmus/r_scripts/complete_merge_time_slices_gexf_created_by_r/create_gexf_complete_merge_budΓ©_era_1536.R | no_license | CKudella/corr_data | R | false | false | 4,292 | r | require(readr)
require(plyr)
require(igraph)
require(rgexf)
# set working directory
getwd()
setwd("../query_results/merge_scripts/complete_merge/")
# read node and edges into dataframe with the name expected by igraph
nodes <- read.csv("time_slice_1536_complete_merge_budΓ©_and_era_correspondents.csv", fileEncoding="UTF-8")
links <- read.csv("time_slice_1536_complete_merge_budΓ©_and_era_letters_corr_as_nodes.csv", fileEncoding="UTF-8")[ ,c('Source', 'Target')]
setwd("../../")
getwd()
mutcorr <- read.csv("./intersection_overview/id_and_names_of_mut_corr_era_budΓ©.csv", fileEncoding="UTF-8")
# add colour for all correspondents
nodes$colour <- "#525252"
# add colour column for mutual correspondents t
nodes$colour <- ifelse(nodes$Id %in% mutcorr$correspondents_id, as.character("#C3161F"), nodes$colour)
#assign specific colour for erasmus
nodes$colour <- ifelse(nodes$Id == "17c580aa-3ba7-4851-8f26-9b3a0ebeadbf", as.character("#3C93AF"), nodes$colour)
#assign specific colour for budΓ©
nodes$colour <- ifelse(nodes$Id == "c0b89c75-45b8-4b04-bfd7-25bfe9ed040b", as.character("#D5AB5B"), nodes$colour)
#assign edge weight
links$weight <- 1
# create igraph object
net <- graph_from_data_frame(d=links, vertices=nodes, directed=T)
# conduct edge bundling (sum edge weights)
net2 <- simplify(net, remove.multiple = TRUE, edge.attr.comb=list(weight="sum","ignore"))
# calculate degree for all nodes
degAll <- degree(net2, v = V(net2), mode = "all")
# calculate weighted degree for all nodes
weightDegAll <- strength(net2, vids = V(net2), mode = "all",
loops = TRUE)
# add new node and edge attributes based on the calculated properties, add
net2 <- set.vertex.attribute(net2, "weightDegAll", index = V(net2), value = weightDegAll)
net2 <- set.vertex.attribute(net2, "degree", index = V(net2), value = degAll)
net2 <- set.vertex.attribute(net2, "colour", index = V(net2), value = nodes$colour)
net2 <- set.edge.attribute(net2, "weight", index = E(net2), value = E(net2)$weight)
#assign edge colour according to source node
edge.start <- ends(net2, es=E(net2), names=F)[,1]
edge.col <- V(net2)$colour[edge.start]
# layout with FR
l <- layout_with_fr(net2, weights=E(net2)$weight)*3.5
# plot graph
plot(net2, layout=l*5, vertex.color=nodes$colour, vertex.size=2, vertex.label=V(net2)$Label, vertex.label.font=2, vertex.label.color="gray40",
vertex.label.cex=.3, edge.arrow.size=.2, edge.width=E(net2)$weight*0.5, edge.color=edge.col, vertex.label.family="sans")
#################
# calculate node coordinates
nodes_coord <- as.data.frame(layout.fruchterman.reingold(net2, weights=E(net2)$weight)*50)
nodes_coord <- cbind(nodes_coord, rep(0, times = nrow(nodes_coord)))
# assign a colour for each node
nodes_col <- V(net2)$colour
# transform nodes into a data frame
nodes_col_df <- as.data.frame(t(col2rgb(nodes_col, alpha = FALSE)))
nodes_col_df <- cbind(nodes_col_df, alpha = rep(1, times = nrow(nodes_col_df)))
# assign visual attributes to nodes (RGBA)
nodes_att_viz <- list(color = nodes_col_df, position = nodes_coord)
# assign a colour for each edge
edges_col <- edge.col
# Transform it into a data frame (we have to transpose it first)
edges_col_df <- as.data.frame(t(col2rgb(edges_col, alpha = FALSE)))
edges_col_df <- cbind(edges_col_df, alpha = rep(1, times = nrow(edges_col_df)))
# assign visual attributes to edges (RGBA)
edges_att_viz <- list(color = edges_col_df)
# create data frames for gexf export
nodes_df <- data.frame(ID = c(1:vcount(net2)), NAME = V(net2)$Label)
edges_df <- as.data.frame(get.edges(net2, c(1:ecount(net2))))
#create a dataframe with node attributes
nodes_att <- data.frame(Degree = V(net2)$degree, colour = as.character(nodes$colour), "Weighted Degree" = V(net2)$weightDegAll)
setwd("../")
getwd()
setwd("./network_data/complete_merge_time_slices_gexf_created_by_r")
# write gexf
era_budΓ©_cmerge_1536 <- write.gexf(nodes = nodes_df, edges = edges_df, edgesWeight = E(net2)$weight, nodesAtt = nodes_att, nodesVizAtt = nodes_att_viz, edgesVizAtt = edges_att_viz, defaultedgetype = "directed", meta = list( creator="Christoph Kudella", description="A graph representing the intersection between Erasmus's and BudΓ©'s networks of correspondence in the year 1536"), output="era_budΓ©_cmerge_1536.gexf") |
#' Spark ML -- Survival Regression
#'
#' Perform survival regression on a Spark DataFrame, using an Accelerated
#' failure time (AFT) model with potentially right-censored data.
#'
#' @template roxlate-ml-x
#' @template roxlate-ml-response
#' @template roxlate-ml-features
#' @template roxlate-ml-intercept
#' @param censor The name of the vector that provides censoring information.
#' This should be a numeric vector, with 0 marking uncensored data, and
#' 1 marking right-censored data.
#' @template roxlate-ml-max-iter
#' @template roxlate-ml-dots
#'
#' @family Spark ML routines
#'
#' @export
ml_survival_regression <- function(x,
response,
features,
intercept = TRUE,
censor = "censor",
max.iter = 100L,
...)
{
df <- spark_dataframe(x)
sc <- spark_connection(df)
df <- ml_prepare_response_features_intercept(df, response, features, intercept)
censor <- ensure_scalar_character(censor)
max.iter <- ensure_scalar_integer(max.iter)
only_model <- ensure_scalar_boolean(list(...)$only_model, default = FALSE)
envir <- new.env(parent = emptyenv())
tdf <- ml_prepare_dataframe(df, features, response, envir = envir)
envir$model <- "org.apache.spark.ml.regression.AFTSurvivalRegression"
rf <- invoke_new(sc, envir$model)
model <- rf %>%
invoke("setMaxIter", max.iter) %>%
invoke("setFeaturesCol", envir$features) %>%
invoke("setLabelCol", envir$response) %>%
invoke("setFitIntercept", as.logical(intercept)) %>%
invoke("setCensorCol", censor)
if (only_model) return(model)
fit <- model %>%
invoke("fit", tdf)
coefficients <- fit %>%
invoke("coefficients") %>%
invoke("toArray")
names(coefficients) <- features
hasIntercept <- invoke(fit, "getFitIntercept")
if (hasIntercept) {
intercept <- invoke(fit, "intercept")
coefficients <- c(coefficients, intercept)
names(coefficients) <- c(features, "(Intercept)")
}
coefficients <- intercept_first(coefficients)
scale <- invoke(fit, "scale")
ml_model("survival_regression", fit,
features = features,
response = response,
intercept = intercept,
coefficients = coefficients,
intercept = intercept,
scale = scale,
model.parameters = as.list(envir)
)
}
#' @export
print.ml_model_survival_regression <- function(x, ...) {
ml_model_print_call(x)
}
| /R/ml_survival_regression.R | permissive | irichgreen/sparklyr | R | false | false | 2,528 | r | #' Spark ML -- Survival Regression
#'
#' Perform survival regression on a Spark DataFrame, using an Accelerated
#' failure time (AFT) model with potentially right-censored data.
#'
#' @template roxlate-ml-x
#' @template roxlate-ml-response
#' @template roxlate-ml-features
#' @template roxlate-ml-intercept
#' @param censor The name of the vector that provides censoring information.
#' This should be a numeric vector, with 0 marking uncensored data, and
#' 1 marking right-censored data.
#' @template roxlate-ml-max-iter
#' @template roxlate-ml-dots
#'
#' @family Spark ML routines
#'
#' @export
ml_survival_regression <- function(x,
response,
features,
intercept = TRUE,
censor = "censor",
max.iter = 100L,
...)
{
df <- spark_dataframe(x)
sc <- spark_connection(df)
df <- ml_prepare_response_features_intercept(df, response, features, intercept)
censor <- ensure_scalar_character(censor)
max.iter <- ensure_scalar_integer(max.iter)
only_model <- ensure_scalar_boolean(list(...)$only_model, default = FALSE)
envir <- new.env(parent = emptyenv())
tdf <- ml_prepare_dataframe(df, features, response, envir = envir)
envir$model <- "org.apache.spark.ml.regression.AFTSurvivalRegression"
rf <- invoke_new(sc, envir$model)
model <- rf %>%
invoke("setMaxIter", max.iter) %>%
invoke("setFeaturesCol", envir$features) %>%
invoke("setLabelCol", envir$response) %>%
invoke("setFitIntercept", as.logical(intercept)) %>%
invoke("setCensorCol", censor)
if (only_model) return(model)
fit <- model %>%
invoke("fit", tdf)
coefficients <- fit %>%
invoke("coefficients") %>%
invoke("toArray")
names(coefficients) <- features
hasIntercept <- invoke(fit, "getFitIntercept")
if (hasIntercept) {
intercept <- invoke(fit, "intercept")
coefficients <- c(coefficients, intercept)
names(coefficients) <- c(features, "(Intercept)")
}
coefficients <- intercept_first(coefficients)
scale <- invoke(fit, "scale")
ml_model("survival_regression", fit,
features = features,
response = response,
intercept = intercept,
coefficients = coefficients,
intercept = intercept,
scale = scale,
model.parameters = as.list(envir)
)
}
#' @export
print.ml_model_survival_regression <- function(x, ...) {
ml_model_print_call(x)
}
|
#
# antlrpdg.R, 17 Jan 20
#
# Data from:
# Hussain Abdullah A. Al-Mutawa
# On the Classification of Cyclic Dependencies in Java Programs
#
# Example from:
# Evidence-based Software Engineering: based on the publicly available data
# Derek M. Jones
#
# TAG Java dependency-cycles
source("ESEUR_config.r")
library("igraph")
library("plyr")
library("jsonlite")
# Increasing default_width does not seem to have any/much effect
plot_layout(3, 2, max_width=ESEUR_default_width+2)
par(oma=OMA_default+c(-1.5, -1.0, -0.2, -0.5))
par(mar=MAR_default+c(-1.5, -3, -0.1, -0.5))
# Remove last name on path (which is assumed to be method name)
remove_last=function(name_str)
{
sub("\\.[a-zA-Z0-9$_]+$", "", name_str)
}
get_src_tgt=function(df)
{
t=c(remove_last(df$src), remove_last(df$tar))
# remove self references
if (t[1] == t[2])
return(NULL)
return(t)
}
plot_PDG=function(file_str)
{
ant=fromJSON(paste0(dir_str, file_str))
from_to=adply(ant$edges, 1, get_src_tgt)
f_t=data.frame(from=from_to$V1, to=from_to$V2)
ant_g=graph.data.frame(unique(f_t), directed=TRUE)
V(ant_g)$label=NA
E(ant_g)$arrow.size=0.3
plot(ant_g, # main=sub("\\.json.xz", "", file_str), # cex.main=2 has no effect!
vertex.frame.color="white")
title(sub("\\.json.xz", "", file_str), cex.main=1.6)
}
dir_str=paste0(ESEUR_dir, "ecosystems/")
top_files=list.files(dir_str)
top_files=top_files[grep("antlr-.*\\.json.xz$", top_files)]
dummy=sapply(top_files, plot_PDG)
| /ecosystems/antlrpdg.R | no_license | sebastianBIanalytics/ESEUR-code-data | R | false | false | 1,445 | r | #
# antlrpdg.R, 17 Jan 20
#
# Data from:
# Hussain Abdullah A. Al-Mutawa
# On the Classification of Cyclic Dependencies in Java Programs
#
# Example from:
# Evidence-based Software Engineering: based on the publicly available data
# Derek M. Jones
#
# TAG Java dependency-cycles
source("ESEUR_config.r")
library("igraph")
library("plyr")
library("jsonlite")
# Increasing default_width does not seem to have any/much effect
plot_layout(3, 2, max_width=ESEUR_default_width+2)
par(oma=OMA_default+c(-1.5, -1.0, -0.2, -0.5))
par(mar=MAR_default+c(-1.5, -3, -0.1, -0.5))
# Remove last name on path (which is assumed to be method name)
remove_last=function(name_str)
{
sub("\\.[a-zA-Z0-9$_]+$", "", name_str)
}
get_src_tgt=function(df)
{
t=c(remove_last(df$src), remove_last(df$tar))
# remove self references
if (t[1] == t[2])
return(NULL)
return(t)
}
plot_PDG=function(file_str)
{
ant=fromJSON(paste0(dir_str, file_str))
from_to=adply(ant$edges, 1, get_src_tgt)
f_t=data.frame(from=from_to$V1, to=from_to$V2)
ant_g=graph.data.frame(unique(f_t), directed=TRUE)
V(ant_g)$label=NA
E(ant_g)$arrow.size=0.3
plot(ant_g, # main=sub("\\.json.xz", "", file_str), # cex.main=2 has no effect!
vertex.frame.color="white")
title(sub("\\.json.xz", "", file_str), cex.main=1.6)
}
dir_str=paste0(ESEUR_dir, "ecosystems/")
top_files=list.files(dir_str)
top_files=top_files[grep("antlr-.*\\.json.xz$", top_files)]
dummy=sapply(top_files, plot_PDG)
|
library(tidyverse)
library(R.utils)
library(DiagrammeR)
library(igraph)
getUnobservedInts <- function(df, desiredResponsesMask, boolLen, str4true) {
# This function first turns the dataframe of parameter-sweep results
# (i.e. observed responses), written in 'pos' and 'neg' format, into
# a vector of binary values (each row is represented by a binary value),
# and then convert the binary values into a list of integers
# which represents the observed integers. After discarding observed integers
# from all possible integers, the funciton returns a list of unobserved integers.
observedInts <-
df %>%
select(., desiredResponsesMask) %>% # select desired responses
apply(., 2, function(x) ifelse(x == str4true, "1", "0")) %>% # pos as 1 & neg as 0
as.data.frame() %>%
apply(., 1, paste0, collapse="") %>% # binary strings representing responses
strtoi(., base = 2L) # convert the binary string into integers
unobservedInts <-
c(0: (2**boolLen-1)) %>% # all possible responses as integers
discard(., . %in% observedInts) # discard observed
return(unobservedInts)
}
getUnobservedInts2 <- function(df, desiredResponsesMask, boolLen) {
# This function is similar to getUnobservedInts, but takes the dataframe
# of parameter-sweep results written in '0' and '1' format.
observedInts <-
df %>%
select(., desiredResponsesMask) %>% # select desired responses
as.data.frame() %>%
apply(., 1, paste0, collapse="") %>% # binary strings representing responses
strtoi(., base = 2L) # convert the binary string into integers
unobservedInts <-
c(0: (2**boolLen-1)) %>% # all possible responses as integers
discard(., . %in% observedInts) # discard observed
return(unobservedInts)
}
getUnobservedBooldf <- function(unobservedInts, desiredResponses){
# The purpose of this function is to turn a list of integers into a
# dataframe in Boolean formats (species responses written as 0s and 1s)
# with an additional output column that meets the requirements of logicopt()
# for Boolean minimization.
unobservedBooldf <-
append((2**length(desiredResponses))-1, unobservedInts) %>%
intToBin(.) %>% # convert the integer back into binary strings
str_split(., "") %>% # chop strings to 0s and 1s
do.call(rbind, .) %>%
as.data.frame() %>%
setNames(., desiredResponses) %>%
slice(., -1) %>%
mutate(., unob = "1") # add an output column to meet the requirements of logicopt()
return(unobservedBooldf)
}
getPCUList <- function(optEqn, str4true, str4flase, desiredResponses){
# This function converts the optimized equations (i.e. results of Boolean minimization)
# into a list of PCUs.
# split the string of equation and get PCUs as a list.
unobservedList <-
str_split(optEqn, " [=] ", simplify = TRUE)[2] %>%
str_split(., " [+] ") %>%
unlist(.) %>%
str_split(., "[*]")
# convert the binary variabe into descriptive strings, e.g.
# RABBITS_ALBATROSSES->"posrabbits_albatrosses" and
# rabbits_albatrosses->"negrabbits_albatrosses"
dictionary <- setNames(
append(paste0(str4true, desiredResponses), paste0(str4flase, desiredResponses)),
append(str_to_upper(desiredResponses), str_to_lower(desiredResponses))
) # create a named vector as dictionary
PCUList_unordered <-
unobservedList %>%
lapply(., function(i) dictionary[i]) %>% # match and replace use dictionary
lapply(., function(i) unname(i)) # unname each list
PCUList <- PCUList_unordered[order(sapply(PCUList_unordered,length))] # sort PCUs
return(PCUList) # show the PCUList
}
get_edgelist_singleAnte <- function(PCUList) {
# Get the edgelist for the simplest kind of implication network,
# where every logical implication statement in its single-antecedent form is included.
andCounter = 0
orCounter = 0
edgesList = list()
for (PCU in PCUList[]){
for (p in PCU[]){
if (length(PCU) == 1) {
ante = 'True'
notpp = ifelse(grepl('pos', p), str_replace(p, 'pos', 'neg'),
str_replace(p, 'neg', 'pos'))
cons = notpp
} else {
ante = p
cons = NULL
qs = PCU[!PCU %in% p]
for (q in qs) {
notqq = ifelse(grepl('pos', q), str_replace(q, 'pos', 'neg'),
str_replace(q, 'neg', 'pos'))
cons = append(cons, notqq)
}
}
# Create the edges list
if (length(cons) == 1) {
edgesList[[length(edgesList)+1]] = append(ante, cons)
} else {
orNode = paste('or', as.character(orCounter), sep = "")
edgesList[[length(edgesList)+1]] = append(ante, orNode)
orCounter = orCounter + 1
for (c in cons) {
edgesList[[length(edgesList)+1]] = append(orNode, c)
}
}
}
}
return(edgesList)
}
get_edgelist_certainAnte <- function(PCUList, alwaysAnteList) {
# Get the edgelist for the implication network where certain responses are
# always treated as the antecedent (alwaysAnteList)
andCounter = 0
orCounter = 0
edgesList = list()
for (PCU in PCUList[]){
# split into a list of antecedents and consequents
anteList = list()
consList_raw = list()
for(p in PCU[]) {
ifelse(p %in% alwaysAnteList, anteList[[length(anteList)+1]] <- p,
consList_raw[[length(consList_raw)+1]] <- p)
}
# The consequents need to be negated
consList = list()
for (q in consList_raw) {
notqq = ifelse(grepl('pos', q), str_replace(q, 'pos', 'neg'),
str_replace(q, 'neg', 'pos'))
consList = append(consList, notqq)
}
# if there are no antecedents, the consequents will be attached to the True node
if (length(anteList) == 0) {anteList = append(anteList, 'True')}
# if there are no consequents, the antecedents will be attached to the False node
if (length(consList) == 0) {consList = append(consList, 'False')}
# identify the upper and lower node, which depends on how many antecedents and consequents
# do upper and antecedents
if (length(anteList) < 2) {
upperNode <- anteList[[1]]
} else {
# the upper node is an And node
andCounter = andCounter + 1
upperNode <- paste('and', as.character(andCounter), sep = "")
# and every antecedent has an edge with this And node
for (p in anteList) {
edgesList[[length(edgesList)+1]] = append(p, upperNode)
}
}
# do lower and consequents
if (length(consList) < 2) {
lowerNode = consList[[1]]
} else {
# the lower node is an Or node
orCounter = orCounter + 1
lowerNode <- paste('or', as.character(orCounter), sep = "")
# and every consequent has an edge with this Or node
for (q in consList) {
edgesList[[length(edgesList)+1]] = append(lowerNode, q)
}
}
# link the upper and lower node
edgesList[[length(edgesList)+1]] = append(upperNode, lowerNode)
}
return(edgesList)
}
draw_implication_network <- function(edgesList, niceNames){
# Draw the implication nework from the edgelist get from the funtion
# get_edgelist_singleAnte() or get_edgelist_certainAnte().
controlSymbol = '↓ ';
# Use edgelist to create an igraph object and get the nodelist from the igraph object
edges_list <- do.call(rbind, edgesList)
nodes_list <-
edges_list %>% graph_from_edgelist(., directed = TRUE) %>%
get.vertex.attribute(.) %>% get("name", .)
# set attributes for response nodes
nodes_df_resp <-
# get response nodes
data.frame(names = nodes_list) %>%
mutate_if(is.factor, as.character) %>%
filter(grepl('pos|neg', names)) %>%
# set sign, fillcolor and node shape
mutate(respSign = case_when(grepl('pos', names) ~ "+", TRUE ~ "‒"),
fillcolor = case_when(grepl('pos', names) ~ "white", TRUE ~ "gray"),
shape = "box") %>%
# separate the strings into two parts: control and response species
mutate(dropSign = substring(names, 4)) %>%
separate(dropSign, c("contSpp", "respSpp"), "_")
# check if niceNames are defined
if (hasArg(niceNames)){
contSppNiceName = unname(niceNames[nodes_df_resp$contSpp])
respSppNiceName = unname(niceNames[nodes_df_resp$respSpp])
} else {
contSppNiceName = nodes_df_resp$contSpp
respSppNiceName = nodes_df_resp$respSpp
}
# set labels for response nodes
nodes_df_resp <- nodes_df_resp %>%
mutate(contSppNiceName = contSppNiceName,
respSppNiceName = respSppNiceName) %>%
mutate(label = paste0('< <font point-size="10">', controlSymbol, contSppNiceName, '</font>',
'<br align="left"/> ', respSppNiceName, ' ', respSign, ' >')) %>%
# select node attribute columns
select(names, shape, fillcolor, label)
# set attributes for Boolean nodes and
# combine response node dataframe and Boolean node dataframe into one dataframe
nodes_df <-
# get Boolean nodes
data.frame(names = nodes_list) %>%
mutate_if(is.factor, as.character) %>%
filter(!grepl('pos|neg', names)) %>%
# set attributes for Boolean nodes
mutate(shape = "circle",
fillcolor = "white",
label = case_when(grepl('or', names) ~ "or",
grepl('and', names) ~ "\"&\"", #NOTE!!!!!!!!!
grepl('False', names) ~ "False",
TRUE ~ "True")) %>%
# combine the two dataframe
bind_rows(., nodes_df_resp)
## construct NDF and EDF for DiagrammeR
ndf <- create_node_df(
n = nrow(nodes_df),
names = nodes_df[, "names"],
label = nodes_df[ , "label"],
shape = nodes_df[, "shape"],
fillcolor = nodes_df[, "fillcolor"])
edges_df <- edges_list %>%
as.data.frame() %>% setNames(., c("labelfrom", "labelto")) %>%
mutate_if(is.factor, as.character) %>%
left_join(., ndf[, c("id", "names")], by = c("labelfrom"="names")) %>% rename(from = id) %>%
left_join(., ndf[, c("id", "names")], by = c("labelto"="names")) %>% rename(to = id)
edf <- create_edge_df(
from = edges_df[ , "from"],
to = edges_df[ , "to"])
G <- create_graph(
nodes_df = ndf,
edges_df = edf,
directed = TRUE,
attr_theme = NULL) %>%
add_global_graph_attrs(
attr = "style",
value = '\"rounded, filled\"',
attr_type = "node") %>%
add_global_graph_attrs(
attr = "width",
value = 0,
attr_type = "node") %>%
add_global_graph_attrs(
attr = "margin",
value = 0,
attr_type = "node")
dot <- gsub("\\'","",generate_dot(G))
DiagrammeR(diagram = dot, type = "grViz")
}
| /qualmodfunc/findPCU.R | no_license | yhan178/qualitative-modeling-r | R | false | false | 12,226 | r | library(tidyverse)
library(R.utils)
library(DiagrammeR)
library(igraph)
getUnobservedInts <- function(df, desiredResponsesMask, boolLen, str4true) {
# This function first turns the dataframe of parameter-sweep results
# (i.e. observed responses), written in 'pos' and 'neg' format, into
# a vector of binary values (each row is represented by a binary value),
# and then convert the binary values into a list of integers
# which represents the observed integers. After discarding observed integers
# from all possible integers, the funciton returns a list of unobserved integers.
observedInts <-
df %>%
select(., desiredResponsesMask) %>% # select desired responses
apply(., 2, function(x) ifelse(x == str4true, "1", "0")) %>% # pos as 1 & neg as 0
as.data.frame() %>%
apply(., 1, paste0, collapse="") %>% # binary strings representing responses
strtoi(., base = 2L) # convert the binary string into integers
unobservedInts <-
c(0: (2**boolLen-1)) %>% # all possible responses as integers
discard(., . %in% observedInts) # discard observed
return(unobservedInts)
}
getUnobservedInts2 <- function(df, desiredResponsesMask, boolLen) {
# This function is similar to getUnobservedInts, but takes the dataframe
# of parameter-sweep results written in '0' and '1' format.
observedInts <-
df %>%
select(., desiredResponsesMask) %>% # select desired responses
as.data.frame() %>%
apply(., 1, paste0, collapse="") %>% # binary strings representing responses
strtoi(., base = 2L) # convert the binary string into integers
unobservedInts <-
c(0: (2**boolLen-1)) %>% # all possible responses as integers
discard(., . %in% observedInts) # discard observed
return(unobservedInts)
}
getUnobservedBooldf <- function(unobservedInts, desiredResponses){
# The purpose of this function is to turn a list of integers into a
# dataframe in Boolean formats (species responses written as 0s and 1s)
# with an additional output column that meets the requirements of logicopt()
# for Boolean minimization.
unobservedBooldf <-
append((2**length(desiredResponses))-1, unobservedInts) %>%
intToBin(.) %>% # convert the integer back into binary strings
str_split(., "") %>% # chop strings to 0s and 1s
do.call(rbind, .) %>%
as.data.frame() %>%
setNames(., desiredResponses) %>%
slice(., -1) %>%
mutate(., unob = "1") # add an output column to meet the requirements of logicopt()
return(unobservedBooldf)
}
getPCUList <- function(optEqn, str4true, str4flase, desiredResponses){
# This function converts the optimized equations (i.e. results of Boolean minimization)
# into a list of PCUs.
# split the string of equation and get PCUs as a list.
unobservedList <-
str_split(optEqn, " [=] ", simplify = TRUE)[2] %>%
str_split(., " [+] ") %>%
unlist(.) %>%
str_split(., "[*]")
# convert the binary variabe into descriptive strings, e.g.
# RABBITS_ALBATROSSES->"posrabbits_albatrosses" and
# rabbits_albatrosses->"negrabbits_albatrosses"
dictionary <- setNames(
append(paste0(str4true, desiredResponses), paste0(str4flase, desiredResponses)),
append(str_to_upper(desiredResponses), str_to_lower(desiredResponses))
) # create a named vector as dictionary
PCUList_unordered <-
unobservedList %>%
lapply(., function(i) dictionary[i]) %>% # match and replace use dictionary
lapply(., function(i) unname(i)) # unname each list
PCUList <- PCUList_unordered[order(sapply(PCUList_unordered,length))] # sort PCUs
return(PCUList) # show the PCUList
}
get_edgelist_singleAnte <- function(PCUList) {
# Get the edgelist for the simplest kind of implication network,
# where every logical implication statement in its single-antecedent form is included.
andCounter = 0
orCounter = 0
edgesList = list()
for (PCU in PCUList[]){
for (p in PCU[]){
if (length(PCU) == 1) {
ante = 'True'
notpp = ifelse(grepl('pos', p), str_replace(p, 'pos', 'neg'),
str_replace(p, 'neg', 'pos'))
cons = notpp
} else {
ante = p
cons = NULL
qs = PCU[!PCU %in% p]
for (q in qs) {
notqq = ifelse(grepl('pos', q), str_replace(q, 'pos', 'neg'),
str_replace(q, 'neg', 'pos'))
cons = append(cons, notqq)
}
}
# Create the edges list
if (length(cons) == 1) {
edgesList[[length(edgesList)+1]] = append(ante, cons)
} else {
orNode = paste('or', as.character(orCounter), sep = "")
edgesList[[length(edgesList)+1]] = append(ante, orNode)
orCounter = orCounter + 1
for (c in cons) {
edgesList[[length(edgesList)+1]] = append(orNode, c)
}
}
}
}
return(edgesList)
}
get_edgelist_certainAnte <- function(PCUList, alwaysAnteList) {
# Get the edgelist for the implication network where certain responses are
# always treated as the antecedent (alwaysAnteList)
andCounter = 0
orCounter = 0
edgesList = list()
for (PCU in PCUList[]){
# split into a list of antecedents and consequents
anteList = list()
consList_raw = list()
for(p in PCU[]) {
ifelse(p %in% alwaysAnteList, anteList[[length(anteList)+1]] <- p,
consList_raw[[length(consList_raw)+1]] <- p)
}
# The consequents need to be negated
consList = list()
for (q in consList_raw) {
notqq = ifelse(grepl('pos', q), str_replace(q, 'pos', 'neg'),
str_replace(q, 'neg', 'pos'))
consList = append(consList, notqq)
}
# if there are no antecedents, the consequents will be attached to the True node
if (length(anteList) == 0) {anteList = append(anteList, 'True')}
# if there are no consequents, the antecedents will be attached to the False node
if (length(consList) == 0) {consList = append(consList, 'False')}
# identify the upper and lower node, which depends on how many antecedents and consequents
# do upper and antecedents
if (length(anteList) < 2) {
upperNode <- anteList[[1]]
} else {
# the upper node is an And node
andCounter = andCounter + 1
upperNode <- paste('and', as.character(andCounter), sep = "")
# and every antecedent has an edge with this And node
for (p in anteList) {
edgesList[[length(edgesList)+1]] = append(p, upperNode)
}
}
# do lower and consequents
if (length(consList) < 2) {
lowerNode = consList[[1]]
} else {
# the lower node is an Or node
orCounter = orCounter + 1
lowerNode <- paste('or', as.character(orCounter), sep = "")
# and every consequent has an edge with this Or node
for (q in consList) {
edgesList[[length(edgesList)+1]] = append(lowerNode, q)
}
}
# link the upper and lower node
edgesList[[length(edgesList)+1]] = append(upperNode, lowerNode)
}
return(edgesList)
}
draw_implication_network <- function(edgesList, niceNames){
# Draw the implication nework from the edgelist get from the funtion
# get_edgelist_singleAnte() or get_edgelist_certainAnte().
controlSymbol = '↓ ';
# Use edgelist to create an igraph object and get the nodelist from the igraph object
edges_list <- do.call(rbind, edgesList)
nodes_list <-
edges_list %>% graph_from_edgelist(., directed = TRUE) %>%
get.vertex.attribute(.) %>% get("name", .)
# set attributes for response nodes
nodes_df_resp <-
# get response nodes
data.frame(names = nodes_list) %>%
mutate_if(is.factor, as.character) %>%
filter(grepl('pos|neg', names)) %>%
# set sign, fillcolor and node shape
mutate(respSign = case_when(grepl('pos', names) ~ "+", TRUE ~ "‒"),
fillcolor = case_when(grepl('pos', names) ~ "white", TRUE ~ "gray"),
shape = "box") %>%
# separate the strings into two parts: control and response species
mutate(dropSign = substring(names, 4)) %>%
separate(dropSign, c("contSpp", "respSpp"), "_")
# check if niceNames are defined
if (hasArg(niceNames)){
contSppNiceName = unname(niceNames[nodes_df_resp$contSpp])
respSppNiceName = unname(niceNames[nodes_df_resp$respSpp])
} else {
contSppNiceName = nodes_df_resp$contSpp
respSppNiceName = nodes_df_resp$respSpp
}
# set labels for response nodes
nodes_df_resp <- nodes_df_resp %>%
mutate(contSppNiceName = contSppNiceName,
respSppNiceName = respSppNiceName) %>%
mutate(label = paste0('< <font point-size="10">', controlSymbol, contSppNiceName, '</font>',
'<br align="left"/> ', respSppNiceName, ' ', respSign, ' >')) %>%
# select node attribute columns
select(names, shape, fillcolor, label)
# set attributes for Boolean nodes and
# combine response node dataframe and Boolean node dataframe into one dataframe
nodes_df <-
# get Boolean nodes
data.frame(names = nodes_list) %>%
mutate_if(is.factor, as.character) %>%
filter(!grepl('pos|neg', names)) %>%
# set attributes for Boolean nodes
mutate(shape = "circle",
fillcolor = "white",
label = case_when(grepl('or', names) ~ "or",
grepl('and', names) ~ "\"&\"", #NOTE!!!!!!!!!
grepl('False', names) ~ "False",
TRUE ~ "True")) %>%
# combine the two dataframe
bind_rows(., nodes_df_resp)
## construct NDF and EDF for DiagrammeR
ndf <- create_node_df(
n = nrow(nodes_df),
names = nodes_df[, "names"],
label = nodes_df[ , "label"],
shape = nodes_df[, "shape"],
fillcolor = nodes_df[, "fillcolor"])
edges_df <- edges_list %>%
as.data.frame() %>% setNames(., c("labelfrom", "labelto")) %>%
mutate_if(is.factor, as.character) %>%
left_join(., ndf[, c("id", "names")], by = c("labelfrom"="names")) %>% rename(from = id) %>%
left_join(., ndf[, c("id", "names")], by = c("labelto"="names")) %>% rename(to = id)
edf <- create_edge_df(
from = edges_df[ , "from"],
to = edges_df[ , "to"])
G <- create_graph(
nodes_df = ndf,
edges_df = edf,
directed = TRUE,
attr_theme = NULL) %>%
add_global_graph_attrs(
attr = "style",
value = '\"rounded, filled\"',
attr_type = "node") %>%
add_global_graph_attrs(
attr = "width",
value = 0,
attr_type = "node") %>%
add_global_graph_attrs(
attr = "margin",
value = 0,
attr_type = "node")
dot <- gsub("\\'","",generate_dot(G))
DiagrammeR(diagram = dot, type = "grViz")
}
|
#' @name osrmTable
#' @title Get Travel Time Matrices Between Points
#' @description Build and send OSRM API queries to get travel time matrices
#' between points. This function interfaces the \emph{table} OSRM service.
#' @param loc a data frame containing 3 fields: points identifiers, longitudes
#' and latitudes (WGS84). It can also be a SpatialPointsDataFrame, a
#' SpatialPolygonsDataFrame or an sf object. If so, row names are used as identifiers.
#' If loc parameter is used, all pair-wise distances are computed.
#' @param src a data frame containing origin points identifiers, longitudes
#' and latitudes (WGS84). It can also be a SpatialPointsDataFrame, a
#' SpatialPolygonsDataFrame or an sf object. If so, row names are used as identifiers.
#' If dst and src parameters are used, only pairs between scr/dst are computed.
#' @param dst a data frame containing destination points identifiers, longitudes
#' and latitudes (WGS84). It can also be a SpatialPointsDataFrame a
#' SpatialPolygonsDataFrame or an sf object. If so, row names are used as identifiers.
#' @param measure a character indicating what measures are calculated. It can
#' be "duration" (in minutes), "distance" (meters), or both c('duration',
#' 'distance'). The demo server only allows "duration".
#' @param exclude pass an optional "exclude" request option to the OSRM API.
#' @param gepaf a boolean indicating if coordinates are sent encoded with the
#' google encoded algorithm format (TRUE) or not (FALSE). Must be FALSE if using
#' the public OSRM API.
#' @param osrm.server the base URL of the routing server.
#' getOption("osrm.server") by default.
#' @param osrm.profile the routing profile to use, e.g. "car", "bike" or "foot"
#' (when using the routing.openstreetmap.de test server).
#' getOption("osrm.profile") by default.
#' @return A list containing 3 data frames is returned.
#' durations is the matrix of travel times (in minutes),
#' sources and destinations are the coordinates of
#' the origin and destination points actually used to compute the travel
#' times (WGS84).
#' @details If loc, src or dst are data frames we assume that the 3 first
#' columns of the data frame are: identifiers, longitudes and latitudes.
#' @note
#' If you want to get a large number of distances make sure to set the
#' "max-table-size" argument (Max. locations supported in table) of the OSRM
#' server accordingly.
#' @seealso \link{osrmIsochrone}
#' @importFrom sf st_as_sf
#' @examples
#' \dontrun{
#' # Load data
#' data("berlin")
#'
#' # Inputs are data frames
#' # Travel time matrix
#' distA <- osrmTable(loc = apotheke.df[1:50, c("id","lon","lat")])
#' # First 5 rows and columns
#' distA$durations[1:5,1:5]
#'
#' # Travel time matrix with different sets of origins and destinations
#' distA2 <- osrmTable(src = apotheke.df[1:10,c("id","lon","lat")],
#' dst = apotheke.df[11:20,c("id","lon","lat")])
#' # First 5 rows and columns
#' distA2$durations[1:5,1:5]
#'
#' # Inputs are sf points
#' distA3 <- osrmTable(loc = apotheke.sf[1:10,])
#' # First 5 rows and columns
#' distA3$durations[1:5,1:5]
#'
#' # Travel time matrix with different sets of origins and destinations
#' distA4 <- osrmTable(src = apotheke.sf[1:10,], dst = apotheke.sf[11:20,])
#' # First 5 rows and columns
#' distA4$durations[1:5,1:5]
#' }
#' @export
osrmTable <- function(loc, src = NULL, dst = NULL, exclude = NULL,
gepaf = FALSE, measure="duration",
osrm.server = getOption("osrm.server"),
osrm.profile = getOption("osrm.profile")){
if(osrm.server == "https://routing.openstreetmap.de/") {
osrm.server = paste0(osrm.server, "routed-", osrm.profile, "/")
osrm.profile = "driving"
}
tryCatch({
# input mgmt
if (is.null(src)){
if(methods::is(loc,"Spatial")){
loc <- st_as_sf(x = loc)
}
if(testSf(loc)){
loc <- sfToDf(x = loc)
}
names(loc) <- c("id", "lon", "lat")
src <- loc
dst <- loc
sep <- "?"
req <- tableLoc(loc = loc, gepaf = gepaf, osrm.server = osrm.server,
osrm.profile = osrm.profile)
}else{
if(methods::is(src,"Spatial")){
src <- st_as_sf(x = src)
}
if(testSf(src)){
src <- sfToDf(x = src)
}
if(methods::is(dst,"Spatial")){
dst <- st_as_sf(x = dst)
}
if(testSf(dst)){
dst <- sfToDf(x = dst)
}
names(src) <- c("id", "lon", "lat")
names(dst) <- c("id", "lon", "lat")
# Build the query
loc <- rbind(src, dst)
sep = "&"
req <- paste(tableLoc(loc = loc, gepaf = gepaf, osrm.server = osrm.server,
osrm.profile = osrm.profile),
"?sources=",
paste(0:(nrow(src)-1), collapse = ";"),
"&destinations=",
paste(nrow(src):(nrow(loc)-1), collapse = ";"),
sep="")
}
# exclude mngmnt
if (!is.null(exclude)) {
exclude_str <- paste0(sep,"exclude=", exclude, sep = "")
sep="&"
}else{
exclude_str <- ""
}
# annotation mngmnt
annotations <- paste0(sep, "annotations=", paste0(measure, collapse=','))
# if(getOption("osrm.server") == "http://router.project-osrm.org/"){
# annotations <- ""
# }
# final req
req <- paste0(req, exclude_str, annotations)
# print(req)
req <- utils::URLencode(req)
osrmLimit(nSrc = nrow(src), nDst = nrow(dst), nreq = nchar(req))
# print(req)
# Get the result
bo=0
while(bo!=10){
x = try({
req_handle <- curl::new_handle(verbose = FALSE)
curl::handle_setopt(req_handle, useragent = "osrm_R_package")
resRaw <- curl::curl(req, handle = req_handle)
res <- jsonlite::fromJSON(resRaw)
}, silent = TRUE)
if (class(x)=="try-error") {
Sys.sleep(1)
bo <- bo+1
} else
break
}
# Check results
if(is.null(res$code)){
e <- simpleError(res$message)
stop(e)
}else{
e <- simpleError(paste0(res$code,"\n",res$message))
if(res$code != "Ok"){stop(e)}
}
output <- list()
if(!is.null(res$durations)){
# get the duration table
output$durations <- durTableFormat(res = res, src = src, dst = dst)
}
if(!is.null(res$distances)){
# get the distance table
output$distances <- distTableFormat(res = res, src = src, dst = dst)
}
# get the coordinates
coords <- coordFormat(res = res, src = src, dst = dst)
output$sources <- coords$sources
output$destinations = coords$destinations
return(output)
}, error=function(e) {message("The OSRM server returned an error:\n", e)})
return(NULL)
}
| /R/osrmTable.R | permissive | IsmailTan35/osrm-backend | R | false | false | 6,869 | r | #' @name osrmTable
#' @title Get Travel Time Matrices Between Points
#' @description Build and send OSRM API queries to get travel time matrices
#' between points. This function interfaces the \emph{table} OSRM service.
#' @param loc a data frame containing 3 fields: points identifiers, longitudes
#' and latitudes (WGS84). It can also be a SpatialPointsDataFrame, a
#' SpatialPolygonsDataFrame or an sf object. If so, row names are used as identifiers.
#' If loc parameter is used, all pair-wise distances are computed.
#' @param src a data frame containing origin points identifiers, longitudes
#' and latitudes (WGS84). It can also be a SpatialPointsDataFrame, a
#' SpatialPolygonsDataFrame or an sf object. If so, row names are used as identifiers.
#' If dst and src parameters are used, only pairs between scr/dst are computed.
#' @param dst a data frame containing destination points identifiers, longitudes
#' and latitudes (WGS84). It can also be a SpatialPointsDataFrame a
#' SpatialPolygonsDataFrame or an sf object. If so, row names are used as identifiers.
#' @param measure a character indicating what measures are calculated. It can
#' be "duration" (in minutes), "distance" (meters), or both c('duration',
#' 'distance'). The demo server only allows "duration".
#' @param exclude pass an optional "exclude" request option to the OSRM API.
#' @param gepaf a boolean indicating if coordinates are sent encoded with the
#' google encoded algorithm format (TRUE) or not (FALSE). Must be FALSE if using
#' the public OSRM API.
#' @param osrm.server the base URL of the routing server.
#' getOption("osrm.server") by default.
#' @param osrm.profile the routing profile to use, e.g. "car", "bike" or "foot"
#' (when using the routing.openstreetmap.de test server).
#' getOption("osrm.profile") by default.
#' @return A list containing 3 data frames is returned.
#' durations is the matrix of travel times (in minutes),
#' sources and destinations are the coordinates of
#' the origin and destination points actually used to compute the travel
#' times (WGS84).
#' @details If loc, src or dst are data frames we assume that the 3 first
#' columns of the data frame are: identifiers, longitudes and latitudes.
#' @note
#' If you want to get a large number of distances make sure to set the
#' "max-table-size" argument (Max. locations supported in table) of the OSRM
#' server accordingly.
#' @seealso \link{osrmIsochrone}
#' @importFrom sf st_as_sf
#' @examples
#' \dontrun{
#' # Load data
#' data("berlin")
#'
#' # Inputs are data frames
#' # Travel time matrix
#' distA <- osrmTable(loc = apotheke.df[1:50, c("id","lon","lat")])
#' # First 5 rows and columns
#' distA$durations[1:5,1:5]
#'
#' # Travel time matrix with different sets of origins and destinations
#' distA2 <- osrmTable(src = apotheke.df[1:10,c("id","lon","lat")],
#' dst = apotheke.df[11:20,c("id","lon","lat")])
#' # First 5 rows and columns
#' distA2$durations[1:5,1:5]
#'
#' # Inputs are sf points
#' distA3 <- osrmTable(loc = apotheke.sf[1:10,])
#' # First 5 rows and columns
#' distA3$durations[1:5,1:5]
#'
#' # Travel time matrix with different sets of origins and destinations
#' distA4 <- osrmTable(src = apotheke.sf[1:10,], dst = apotheke.sf[11:20,])
#' # First 5 rows and columns
#' distA4$durations[1:5,1:5]
#' }
#' @export
osrmTable <- function(loc, src = NULL, dst = NULL, exclude = NULL,
gepaf = FALSE, measure="duration",
osrm.server = getOption("osrm.server"),
osrm.profile = getOption("osrm.profile")){
if(osrm.server == "https://routing.openstreetmap.de/") {
osrm.server = paste0(osrm.server, "routed-", osrm.profile, "/")
osrm.profile = "driving"
}
tryCatch({
# input mgmt
if (is.null(src)){
if(methods::is(loc,"Spatial")){
loc <- st_as_sf(x = loc)
}
if(testSf(loc)){
loc <- sfToDf(x = loc)
}
names(loc) <- c("id", "lon", "lat")
src <- loc
dst <- loc
sep <- "?"
req <- tableLoc(loc = loc, gepaf = gepaf, osrm.server = osrm.server,
osrm.profile = osrm.profile)
}else{
if(methods::is(src,"Spatial")){
src <- st_as_sf(x = src)
}
if(testSf(src)){
src <- sfToDf(x = src)
}
if(methods::is(dst,"Spatial")){
dst <- st_as_sf(x = dst)
}
if(testSf(dst)){
dst <- sfToDf(x = dst)
}
names(src) <- c("id", "lon", "lat")
names(dst) <- c("id", "lon", "lat")
# Build the query
loc <- rbind(src, dst)
sep = "&"
req <- paste(tableLoc(loc = loc, gepaf = gepaf, osrm.server = osrm.server,
osrm.profile = osrm.profile),
"?sources=",
paste(0:(nrow(src)-1), collapse = ";"),
"&destinations=",
paste(nrow(src):(nrow(loc)-1), collapse = ";"),
sep="")
}
# exclude mngmnt
if (!is.null(exclude)) {
exclude_str <- paste0(sep,"exclude=", exclude, sep = "")
sep="&"
}else{
exclude_str <- ""
}
# annotation mngmnt
annotations <- paste0(sep, "annotations=", paste0(measure, collapse=','))
# if(getOption("osrm.server") == "http://router.project-osrm.org/"){
# annotations <- ""
# }
# final req
req <- paste0(req, exclude_str, annotations)
# print(req)
req <- utils::URLencode(req)
osrmLimit(nSrc = nrow(src), nDst = nrow(dst), nreq = nchar(req))
# print(req)
# Get the result
bo=0
while(bo!=10){
x = try({
req_handle <- curl::new_handle(verbose = FALSE)
curl::handle_setopt(req_handle, useragent = "osrm_R_package")
resRaw <- curl::curl(req, handle = req_handle)
res <- jsonlite::fromJSON(resRaw)
}, silent = TRUE)
if (class(x)=="try-error") {
Sys.sleep(1)
bo <- bo+1
} else
break
}
# Check results
if(is.null(res$code)){
e <- simpleError(res$message)
stop(e)
}else{
e <- simpleError(paste0(res$code,"\n",res$message))
if(res$code != "Ok"){stop(e)}
}
output <- list()
if(!is.null(res$durations)){
# get the duration table
output$durations <- durTableFormat(res = res, src = src, dst = dst)
}
if(!is.null(res$distances)){
# get the distance table
output$distances <- distTableFormat(res = res, src = src, dst = dst)
}
# get the coordinates
coords <- coordFormat(res = res, src = src, dst = dst)
output$sources <- coords$sources
output$destinations = coords$destinations
return(output)
}, error=function(e) {message("The OSRM server returned an error:\n", e)})
return(NULL)
}
|
% Generated by roxygen2 (4.0.1): do not edit by hand
\name{extract_ILDIS}
\alias{extract_ILDIS}
\title{Extract geographical records data from ILDIS species page html}
\usage{
extract_ILDIS(html_file)
}
\arguments{
\item{html_file}{The path of the html file to process}
}
\description{
Extract geographical records data from ILDIS species page html
}
\details{
The name of the species is extracted from the filename, as generated by \code{\link{download_ILDIS}}.
}
| /man/extract_ILDIS.Rd | permissive | rdinnager/GlobalNfix | R | false | false | 465 | rd | % Generated by roxygen2 (4.0.1): do not edit by hand
\name{extract_ILDIS}
\alias{extract_ILDIS}
\title{Extract geographical records data from ILDIS species page html}
\usage{
extract_ILDIS(html_file)
}
\arguments{
\item{html_file}{The path of the html file to process}
}
\description{
Extract geographical records data from ILDIS species page html
}
\details{
The name of the species is extracted from the filename, as generated by \code{\link{download_ILDIS}}.
}
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/sim_bessel_layers.R
\name{simulate_layered_brownian_bridge_bessel}
\alias{simulate_layered_brownian_bridge_bessel}
\title{Layered Brownian Bridge sampler}
\usage{
simulate_layered_brownian_bridge_bessel(x, y, s, t, a, l, sim_times)
}
\arguments{
\item{x}{start value of Brownian bridge}
\item{y}{end value of Brownian bridge}
\item{s}{start value of Brownian bridge}
\item{t}{end value of Brownian bridge}
\item{a}{vector/sequence of numbers}
\item{l}{integer number denoting Bessel layer, i.e. Brownian bridge is contained in [min(x,y)-a[l], max(x,y)+a[l]]}
\item{sim_times}{vector of real numbers to simulate Bessel bridge}
}
\value{
matrix of the simulated layered Brownian bridge path, first row is points X, second row are corresponding times
}
\description{
This function simulates a layered Brownian Bridge given a Bessel layer, at given times
}
\examples{
# simulate Bessel layer
bes_layer <- simulate_bessel_layer(x = 0, y = 0, s = 0, t = 1, a = seq(0.1, 1.0, 0.1))
# simulate layered Brownian bridge
simulate_layered_brownian_bridge_bessel(x = 0, y = 0, s = 0, t = 1, a = bes_layer$a, l = bes_layer$l, sim_times = seq(0.2, 0.8, 0.2))
}
| /man/simulate_layered_brownian_bridge_bessel.Rd | no_license | rchan26/RlayeredBB | R | false | true | 1,231 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/sim_bessel_layers.R
\name{simulate_layered_brownian_bridge_bessel}
\alias{simulate_layered_brownian_bridge_bessel}
\title{Layered Brownian Bridge sampler}
\usage{
simulate_layered_brownian_bridge_bessel(x, y, s, t, a, l, sim_times)
}
\arguments{
\item{x}{start value of Brownian bridge}
\item{y}{end value of Brownian bridge}
\item{s}{start value of Brownian bridge}
\item{t}{end value of Brownian bridge}
\item{a}{vector/sequence of numbers}
\item{l}{integer number denoting Bessel layer, i.e. Brownian bridge is contained in [min(x,y)-a[l], max(x,y)+a[l]]}
\item{sim_times}{vector of real numbers to simulate Bessel bridge}
}
\value{
matrix of the simulated layered Brownian bridge path, first row is points X, second row are corresponding times
}
\description{
This function simulates a layered Brownian Bridge given a Bessel layer, at given times
}
\examples{
# simulate Bessel layer
bes_layer <- simulate_bessel_layer(x = 0, y = 0, s = 0, t = 1, a = seq(0.1, 1.0, 0.1))
# simulate layered Brownian bridge
simulate_layered_brownian_bridge_bessel(x = 0, y = 0, s = 0, t = 1, a = bes_layer$a, l = bes_layer$l, sim_times = seq(0.2, 0.8, 0.2))
}
|
source("knnPerformanceNoCV.R")
DPI_list = c(100,200,300);
k_list = seq(1,40,1);
GroupNumber = 2;
MemberNumber = 2;
sigma_list = c(1.5,2.5,3.5);
n_list = seq(500,4000,500);
DPI_tests <- list()
i=1
DPI=DPI_list[i]
sigma=sigma_list[i]
DataList = loadSinglePersonsData(DPI,GroupNumber,MemberNumber,sigma)
for(j in 1:10)
{
if(j == 1)
{
data = DataList[[j]];
dataClass = rep(j-1,nrow(DataList[[j]]));
}
else
{
data = rbind(data, DataList[[j]]);
dataClass = append(dataClass, rep(j-1, nrow(DataList[[j]]) ) );
}
}
dataClassF = factor(dataClass)
remove(DataList);
gc();
knnTest=knnPerformanceOnTrainingSetNoCV(data,dataClassF,k_list)
knnTestFilenamePF=paste(c("../data/knnTrainingSetPerformance-",GroupNumber,"-",MemberNumber,"-",DPI,"-",sigma),collapse="")
save(knnTest,file=paste(c(knnTestFilenamePF,".RData"),collapse=""))
load(paste(c(knnTestFilenamePF,".RData"),collapse=""))
acc = vector(mode="double",length=length(k_list))
acclow = vector(mode="double",length=length(k_list))
acchigh = vector(mode="double",length=length(k_list))
for(i in 1:(length(k_list)))
{
acc[i]=knnTest$knnConfusionMatrix[[i]]$overall[["Accuracy"]]
acclow[i]=knnTest$knnConfusionMatrix[[i]]$overall[["AccuracyLower"]]
acchigh[i]=knnTest$knnConfusionMatrix[[i]]$overall[["AccuracyUpper"]]
}
knnTestResults=data.frame(k=k_list,Accuracy=acc,Accuracy95Low=acclow,Accuracy95High=acchigh)
write.csv(x=knnTestResults,file=paste(c(knnTestFilenamePF,".csv"),collapse=""),row.names = FALSE)
| /Rcode/test-k-NN-trainingSetPerformance.R | no_license | madsherlock/SML-F16 | R | false | false | 1,530 | r | source("knnPerformanceNoCV.R")
DPI_list = c(100,200,300);
k_list = seq(1,40,1);
GroupNumber = 2;
MemberNumber = 2;
sigma_list = c(1.5,2.5,3.5);
n_list = seq(500,4000,500);
DPI_tests <- list()
i=1
DPI=DPI_list[i]
sigma=sigma_list[i]
DataList = loadSinglePersonsData(DPI,GroupNumber,MemberNumber,sigma)
for(j in 1:10)
{
if(j == 1)
{
data = DataList[[j]];
dataClass = rep(j-1,nrow(DataList[[j]]));
}
else
{
data = rbind(data, DataList[[j]]);
dataClass = append(dataClass, rep(j-1, nrow(DataList[[j]]) ) );
}
}
dataClassF = factor(dataClass)
remove(DataList);
gc();
knnTest=knnPerformanceOnTrainingSetNoCV(data,dataClassF,k_list)
knnTestFilenamePF=paste(c("../data/knnTrainingSetPerformance-",GroupNumber,"-",MemberNumber,"-",DPI,"-",sigma),collapse="")
save(knnTest,file=paste(c(knnTestFilenamePF,".RData"),collapse=""))
load(paste(c(knnTestFilenamePF,".RData"),collapse=""))
acc = vector(mode="double",length=length(k_list))
acclow = vector(mode="double",length=length(k_list))
acchigh = vector(mode="double",length=length(k_list))
for(i in 1:(length(k_list)))
{
acc[i]=knnTest$knnConfusionMatrix[[i]]$overall[["Accuracy"]]
acclow[i]=knnTest$knnConfusionMatrix[[i]]$overall[["AccuracyLower"]]
acchigh[i]=knnTest$knnConfusionMatrix[[i]]$overall[["AccuracyUpper"]]
}
knnTestResults=data.frame(k=k_list,Accuracy=acc,Accuracy95Low=acclow,Accuracy95High=acchigh)
write.csv(x=knnTestResults,file=paste(c(knnTestFilenamePF,".csv"),collapse=""),row.names = FALSE)
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/ordinal_ridge.R
\name{evaluateRanking}
\alias{evaluateRanking}
\title{Evaluate rank prediction performance}
\usage{
evaluateRanking(scores, labels)
}
\arguments{
\item{scores}{Numeric vector of a score ranking the observations in the predicted order}
\item{labels}{Ordered factor vector of true labels for each observation}
}
\value{
A score between 0 and 1. Higher scores mean better prediction performance
}
\description{
Evaluate rank prediction performance
}
| /man/evaluateRanking.Rd | permissive | ArtemSokolov/ordinalRidge | R | false | true | 542 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/ordinal_ridge.R
\name{evaluateRanking}
\alias{evaluateRanking}
\title{Evaluate rank prediction performance}
\usage{
evaluateRanking(scores, labels)
}
\arguments{
\item{scores}{Numeric vector of a score ranking the observations in the predicted order}
\item{labels}{Ordered factor vector of true labels for each observation}
}
\value{
A score between 0 and 1. Higher scores mean better prediction performance
}
\description{
Evaluate rank prediction performance
}
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/pkg_version.R
\name{pkg_version}
\alias{pkg_version}
\title{Retrieve Package version}
\usage{
pkg_version(pkgs_col)
}
\arguments{
\item{pkgs_col}{Package name.}
}
\value{
A character vector with the package version.
}
\description{
Internal helper function.
}
| /man/pkg_version.Rd | permissive | luisDVA/annotater | R | false | true | 338 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/pkg_version.R
\name{pkg_version}
\alias{pkg_version}
\title{Retrieve Package version}
\usage{
pkg_version(pkgs_col)
}
\arguments{
\item{pkgs_col}{Package name.}
}
\value{
A character vector with the package version.
}
\description{
Internal helper function.
}
|
options( show.error.messages=F, error = function () { cat( geterrmessage(), file=stderr() ); q( "no", 1, F ) } )
# we need that to not crash galaxy with an UTF8 error on German LC settings.
loc <- Sys.setlocale("LC_MESSAGES", "en_US.UTF-8")
suppressPackageStartupMessages({
library(ChIPseeker)
library(GenomicFeatures)
library(rtracklayer)
library(optparse)
})
option_list <- list(
make_option(c("-i","--infile"), type="character", help="Peaks file to be annotated"),
make_option(c("-G","--gtf"), type="character", help="GTF to create TxDb."),
make_option(c("-u","--upstream"), type="integer", help="TSS upstream region"),
make_option(c("-d","--downstream"), type="integer", help="TSS downstream region"),
make_option(c("-F","--flankgeneinfo"), type="logical", help="Add flanking gene info"),
make_option(c("-D","--flankgenedist"), type="integer", help="Flanking gene distance"),
make_option(c("-f","--format"), type="character", help="Output format (interval or tabular)."),
make_option(c("-p","--plots"), type="logical", help="PDF of plots."),
make_option(c("-r","--rdata"), type="logical", help="Output RData file.")
)
parser <- OptionParser(usage = "%prog [options] file", option_list=option_list)
args = parse_args(parser)
peaks = args$infile
gtf = args$gtf
up = args$upstream
down = args$downstream
format = args$format
peaks <- readPeakFile(peaks)
# Make TxDb from GTF
txdb <- makeTxDbFromGFF(gtf, format="gtf")
if (!is.null(args$flankgeneinfo)) {
peakAnno <- annotatePeak(peaks, TxDb=txdb, tssRegion=c(-up, down), addFlankGeneInfo=args$flankgeneinfo, flankDistance=args$flankgenedist)
} else {
peakAnno <- annotatePeak(peaks, TxDb=txdb, tssRegion=c(-up, down))
}
# Add gene name
features <- import(gtf, format="gtf")
ann <- unique(mcols(features)[, c("gene_id", "gene_name")])
res <- as.data.frame(peakAnno)
res <- merge(res, ann, by.x="geneId", by.y="gene_id")
names(res)[names(res) == "gene_name"] <- "geneName"
#Extract metadata cols, 1st is geneId, rest should be from col 7 to end
metacols <- res[, c(7:ncol(res), 1)]
# Convert from 1-based to 0-based format
if (format == "interval") {
metacols <- apply(as.data.frame(metacols), 1, function(col) paste(col, collapse="|"))
resout <- data.frame(Chrom=res$seqnames,
Start=res$start - 1,
End=res$end,
Comment=metacols)
} else {
resout <- data.frame(Chrom=res$seqnames,
Start=res$start - 1,
End=res$end,
metacols)
}
write.table(resout, file="out.tab", sep="\t", row.names=FALSE, quote=FALSE)
if (!is.null(args$plots)) {
pdf("out.pdf", width=14)
plotAnnoPie(peakAnno)
plotAnnoBar(peakAnno)
vennpie(peakAnno)
upsetplot(peakAnno)
plotDistToTSS(peakAnno, title="Distribution of transcription factor-binding loci\nrelative to TSS")
dev.off()
}
## Output RData file
if (!is.null(args$rdata)) {
save.image(file = "ChIPseeker_analysis.RData")
} | /tools/chipseeker/chipseeker.R | no_license | wm75/galaxytools | R | false | false | 3,034 | r | options( show.error.messages=F, error = function () { cat( geterrmessage(), file=stderr() ); q( "no", 1, F ) } )
# we need that to not crash galaxy with an UTF8 error on German LC settings.
loc <- Sys.setlocale("LC_MESSAGES", "en_US.UTF-8")
suppressPackageStartupMessages({
library(ChIPseeker)
library(GenomicFeatures)
library(rtracklayer)
library(optparse)
})
option_list <- list(
make_option(c("-i","--infile"), type="character", help="Peaks file to be annotated"),
make_option(c("-G","--gtf"), type="character", help="GTF to create TxDb."),
make_option(c("-u","--upstream"), type="integer", help="TSS upstream region"),
make_option(c("-d","--downstream"), type="integer", help="TSS downstream region"),
make_option(c("-F","--flankgeneinfo"), type="logical", help="Add flanking gene info"),
make_option(c("-D","--flankgenedist"), type="integer", help="Flanking gene distance"),
make_option(c("-f","--format"), type="character", help="Output format (interval or tabular)."),
make_option(c("-p","--plots"), type="logical", help="PDF of plots."),
make_option(c("-r","--rdata"), type="logical", help="Output RData file.")
)
parser <- OptionParser(usage = "%prog [options] file", option_list=option_list)
args = parse_args(parser)
peaks = args$infile
gtf = args$gtf
up = args$upstream
down = args$downstream
format = args$format
peaks <- readPeakFile(peaks)
# Make TxDb from GTF
txdb <- makeTxDbFromGFF(gtf, format="gtf")
if (!is.null(args$flankgeneinfo)) {
peakAnno <- annotatePeak(peaks, TxDb=txdb, tssRegion=c(-up, down), addFlankGeneInfo=args$flankgeneinfo, flankDistance=args$flankgenedist)
} else {
peakAnno <- annotatePeak(peaks, TxDb=txdb, tssRegion=c(-up, down))
}
# Add gene name
features <- import(gtf, format="gtf")
ann <- unique(mcols(features)[, c("gene_id", "gene_name")])
res <- as.data.frame(peakAnno)
res <- merge(res, ann, by.x="geneId", by.y="gene_id")
names(res)[names(res) == "gene_name"] <- "geneName"
#Extract metadata cols, 1st is geneId, rest should be from col 7 to end
metacols <- res[, c(7:ncol(res), 1)]
# Convert from 1-based to 0-based format
if (format == "interval") {
metacols <- apply(as.data.frame(metacols), 1, function(col) paste(col, collapse="|"))
resout <- data.frame(Chrom=res$seqnames,
Start=res$start - 1,
End=res$end,
Comment=metacols)
} else {
resout <- data.frame(Chrom=res$seqnames,
Start=res$start - 1,
End=res$end,
metacols)
}
write.table(resout, file="out.tab", sep="\t", row.names=FALSE, quote=FALSE)
if (!is.null(args$plots)) {
pdf("out.pdf", width=14)
plotAnnoPie(peakAnno)
plotAnnoBar(peakAnno)
vennpie(peakAnno)
upsetplot(peakAnno)
plotDistToTSS(peakAnno, title="Distribution of transcription factor-binding loci\nrelative to TSS")
dev.off()
}
## Output RData file
if (!is.null(args$rdata)) {
save.image(file = "ChIPseeker_analysis.RData")
} |
## Kaggle - NIPS 2015 Papers
# Packages
library(readr)
library(dplyr)
# Read input data
authors = read_csv("input/Authors.csv")
paperAuthors = read_csv("input/PaperAuthors.csv")
papers = read_csv("input/Papers.csv")
#Look at type of events at NIPS
table(papers$EventType)
#Function to find last name for each row
findLastName = function(fullname){
return(fullname[length(fullname)])
}
##
## Check last names for authors
##
#Check last names
authorNames = strsplit(authors$Name, " ")
lastNames = lapply(authorNames, findLastName)
lastNames = as.character(lastNames)
#Count number of each last name
lastNamesCount = as.data.frame(table(lastNames))
lastNamesCount = lastNamesCount[order(lastNamesCount$Freq),]
##
## Check last name counts for authors linked to papers
##
#Find who wrote papers
publishers = merge(authors, paperAuthors, by.x = "Id", by.y="AuthorId")
#Check last names
authorNames = strsplit(publishers$Name, " ")
lastNamesPub = lapply(authorNames, findLastName)
lastNamesPub = as.character(lastNamesPub)
#Count number of each last name
lastNamesPubCount = as.data.frame(table(lastNamesPub))
lastNamesPubCount = lastNamesPubCount[order(lastNamesPubCount$Freq),]
| /NIPS 2015 Papers/exploratory.R | no_license | Anithaponnuru/Kaggle | R | false | false | 1,188 | r | ## Kaggle - NIPS 2015 Papers
# Packages
library(readr)
library(dplyr)
# Read input data
authors = read_csv("input/Authors.csv")
paperAuthors = read_csv("input/PaperAuthors.csv")
papers = read_csv("input/Papers.csv")
#Look at type of events at NIPS
table(papers$EventType)
#Function to find last name for each row
findLastName = function(fullname){
return(fullname[length(fullname)])
}
##
## Check last names for authors
##
#Check last names
authorNames = strsplit(authors$Name, " ")
lastNames = lapply(authorNames, findLastName)
lastNames = as.character(lastNames)
#Count number of each last name
lastNamesCount = as.data.frame(table(lastNames))
lastNamesCount = lastNamesCount[order(lastNamesCount$Freq),]
##
## Check last name counts for authors linked to papers
##
#Find who wrote papers
publishers = merge(authors, paperAuthors, by.x = "Id", by.y="AuthorId")
#Check last names
authorNames = strsplit(publishers$Name, " ")
lastNamesPub = lapply(authorNames, findLastName)
lastNamesPub = as.character(lastNamesPub)
#Count number of each last name
lastNamesPubCount = as.data.frame(table(lastNamesPub))
lastNamesPubCount = lastNamesPubCount[order(lastNamesPubCount$Freq),]
|
source("2__scripts/1__R/3__Release/RTools.R")
Instal_Required("data.table")
Instal_Required("ade4")
Instal_Required("tree")
Instal_Required("lda")
Instal_Required("ggplot2")
Instal_Required("randomForest")
#setwd(dir = "C:/Users/felix.rougier/Documents/Challenge/DataScienceNet/maif/")
maif_train <- fread("1__data/1__input/Brut_Train.csv", header=T)
maif_test <- fread("1__data/1__input/Brut_Test.csv", header=T)
# analyse de la donn?es crm
# peut-on l'utiliser pour obtenir facilement le prix d'achat de base :
# distribution de CRM:
d <- density(maif_train$crm)
plot(d, col='red')
lines(density(maif_test$crm), col='blue')
hist(maif_train$crm, freq=F, ylim=c(0,0.09), main="R?partition de CRM")
lines(density(maif_train$crm), col='red')
t <- round(100*table(maif_train$crm)/nrow(maif_train),2)
t
summary(maif_train$crm)
# va jusqu'? 270 mais cas tr?s particuliers :
# moins de 0.01% de valeurs au dessus de 195
# comparaison des prix bruts et des prix finaux
# prime totale
d <- maif_train[,.(crm,prime_tot_ttc)]
d
plot(d$crm,d$prime_tot_ttc)
# prime brute avant crm
d[,prime_brute:=100*prime_tot_ttc/crm]
plot(d$crm,d$prime_brute)
# calcul des moyennes
d[,mean_prime_tot:=mean(prime_tot_ttc), by='crm']
d[,mean_prime_brute:=mean(prime_brute), by='crm']
# repr?sentation graphique
d2 <- unique(d[,.(crm,mean_prime_tot,mean_prime_brute)])
setkey(d2,crm)
plot(d2$crm, d2$mean_prime_tot , type='b', col='blue', ylim=c(0,1300))
points(d2$crm, d2$mean_prime_brute , type='b', col='red')
# test
d2[,prime_mean:=crm*mean_prime_tot/100]
d2
points(d2$crm,d2$prime_mean, type='b', col='green')
# peut-?tre que le facteur d'application du crm n'est pas exactement prix*crm/100
# on va faire une r?gression lin?aire pour d?terminer le rapport entre le crm e le prix
# sur les moyennes des prix (non pond?r?)
mod1 <- lm(d2$mean_prime_tot~d2$crm)
summary(mod1)
# sur l'ensemble des prix
mod2 <- lm(maif_train$prime_tot_ttc~maif_train$crm)
summary(mod2)
#
mod3 <- lm( rep(m,300000) ~ maif_train$prime_tot_ttc * maif_train$crm -1 )
summary(mod3)
# ne fontionne pas
# on va consid?rer que le crm s'applique bien comme un bonus malus et qu'on retrouve le prix brut de :
# Prix de Base = prime_totale_ttc * 100 / CRM
| /2__scripts/1__R/1__AD/AD_crm.R | no_license | romainjouen/KaggleMAIF | R | false | false | 2,251 | r | source("2__scripts/1__R/3__Release/RTools.R")
Instal_Required("data.table")
Instal_Required("ade4")
Instal_Required("tree")
Instal_Required("lda")
Instal_Required("ggplot2")
Instal_Required("randomForest")
#setwd(dir = "C:/Users/felix.rougier/Documents/Challenge/DataScienceNet/maif/")
maif_train <- fread("1__data/1__input/Brut_Train.csv", header=T)
maif_test <- fread("1__data/1__input/Brut_Test.csv", header=T)
# analyse de la donn?es crm
# peut-on l'utiliser pour obtenir facilement le prix d'achat de base :
# distribution de CRM:
d <- density(maif_train$crm)
plot(d, col='red')
lines(density(maif_test$crm), col='blue')
hist(maif_train$crm, freq=F, ylim=c(0,0.09), main="R?partition de CRM")
lines(density(maif_train$crm), col='red')
t <- round(100*table(maif_train$crm)/nrow(maif_train),2)
t
summary(maif_train$crm)
# va jusqu'? 270 mais cas tr?s particuliers :
# moins de 0.01% de valeurs au dessus de 195
# comparaison des prix bruts et des prix finaux
# prime totale
d <- maif_train[,.(crm,prime_tot_ttc)]
d
plot(d$crm,d$prime_tot_ttc)
# prime brute avant crm
d[,prime_brute:=100*prime_tot_ttc/crm]
plot(d$crm,d$prime_brute)
# calcul des moyennes
d[,mean_prime_tot:=mean(prime_tot_ttc), by='crm']
d[,mean_prime_brute:=mean(prime_brute), by='crm']
# repr?sentation graphique
d2 <- unique(d[,.(crm,mean_prime_tot,mean_prime_brute)])
setkey(d2,crm)
plot(d2$crm, d2$mean_prime_tot , type='b', col='blue', ylim=c(0,1300))
points(d2$crm, d2$mean_prime_brute , type='b', col='red')
# test
d2[,prime_mean:=crm*mean_prime_tot/100]
d2
points(d2$crm,d2$prime_mean, type='b', col='green')
# peut-?tre que le facteur d'application du crm n'est pas exactement prix*crm/100
# on va faire une r?gression lin?aire pour d?terminer le rapport entre le crm e le prix
# sur les moyennes des prix (non pond?r?)
mod1 <- lm(d2$mean_prime_tot~d2$crm)
summary(mod1)
# sur l'ensemble des prix
mod2 <- lm(maif_train$prime_tot_ttc~maif_train$crm)
summary(mod2)
#
mod3 <- lm( rep(m,300000) ~ maif_train$prime_tot_ttc * maif_train$crm -1 )
summary(mod3)
# ne fontionne pas
# on va consid?rer que le crm s'applique bien comme un bonus malus et qu'on retrouve le prix brut de :
# Prix de Base = prime_totale_ttc * 100 / CRM
|
source('/Users/zeynepenkavi/Dropbox/PoldrackLab/SRO_Retest_Analyses/code/figure_scripts/figure_res_wrapper.R')
if(!exists('rel_df')){
source('/Users/zeynepenkavi/Dropbox/PoldrackLab/SRO_Retest_Analyses/code/workspace_scripts/subject_data.R')
source('/Users/zeynepenkavi/Dropbox/PoldrackLab/SRO_Retest_Analyses/code/helper_functions/make_rel_df.R')
rel_df = make_rel_df(t1_df = test_data, t2_df = retest_data, metrics = c('spearman', 'icc', 'pearson', 'var_breakdown', 'partial_eta', 'sem'))
rel_df$task = 'task'
rel_df[grep('survey', rel_df$dv), 'task'] = 'survey'
rel_df[grep('holt', rel_df$dv), 'task'] = "task"
rel_df = rel_df %>%
select(dv, task, spearman, icc, pearson, partial_eta, sem, var_subs, var_ind, var_resid)
}
tmp = rel_df %>%
mutate(var_subs_pct = var_subs/(var_subs+var_ind+var_resid)*100,
var_ind_pct = var_ind/(var_subs+var_ind+var_resid)*100,
var_resid_pct = var_resid/(var_subs+var_ind+var_resid)*100) %>%
select(dv, task, var_subs_pct, var_ind_pct, var_resid_pct) %>%
mutate(dv = factor(dv, levels = dv[order(task)])) %>%
separate(dv, c("task_group", "var"), sep="\\.",remove=FALSE,extra="merge") %>%
mutate(task_group = factor(task_group, levels = task_group[order(task)])) %>%
arrange(task_group, var_subs_pct) %>%
mutate(rank = row_number()) %>%
arrange(task, task_group, rank) %>%
gather(key, value, -dv, -task_group, -var, -task, -rank) %>%
ungroup()%>%
mutate(task_group = gsub("_", " ", task_group),
var = gsub("_", " ", var)) %>%
mutate(task_group = ifelse(task_group == "psychological refractory period two choices", "psychological refractory period", ifelse(task_group == "angling risk task always sunny", "angling risk task",task_group))) %>%
mutate(task_group = gsub("survey", "", task_group)) %>%
filter(task=="task",
!grepl("EZ|hddm", dv))%>%
arrange(task_group, rank)
labels = tmp %>%
distinct(dv, .keep_all=T)
p1 <- tmp %>%
ggplot(aes(x=factor(rank), y=value, fill=factor(key, levels = c("var_resid_pct", "var_ind_pct", "var_subs_pct"))))+
geom_bar(stat='identity', alpha = 0.75, color='#00BFC4')+
scale_x_discrete(breaks = labels$rank,
labels = labels$var)+
coord_flip()+
facet_grid(task_group~., switch = "y", scales = "free_y", space = "free_y") +
theme(panel.spacing = unit(0.5, "lines"),
strip.placement = "outside",
strip.text.y = element_text(angle=180),
panel.background = element_rect(fill = NA),
panel.grid.major = element_line(colour = "grey85"),
legend.position = 'bottom')+
theme(legend.title = element_blank())+
scale_fill_manual(breaks = c("var_subs_pct", "var_ind_pct", "var_resid_pct"),
labels = c("Variance between individuals",
"Variance between sessions",
"Error variance"),
values=c("grey65", "grey45", "grey25"))+
ylab("")+
xlab("")
tmp = rel_df %>%
mutate(var_subs_pct = var_subs/(var_subs+var_ind+var_resid)*100,
var_ind_pct = var_ind/(var_subs+var_ind+var_resid)*100,
var_resid_pct = var_resid/(var_subs+var_ind+var_resid)*100) %>%
select(dv, task, var_subs_pct, var_ind_pct, var_resid_pct) %>%
mutate(dv = factor(dv, levels = dv[order(task)])) %>%
separate(dv, c("task_group", "var"), sep="\\.",remove=FALSE,extra="merge") %>%
mutate(task_group = factor(task_group, levels = task_group[order(task)])) %>%
arrange(task_group, var_subs_pct) %>%
mutate(rank = row_number()) %>%
arrange(task, task_group, rank) %>%
gather(key, value, -dv, -task_group, -var, -task, -rank) %>%
ungroup()%>%
mutate(task_group = gsub("_", " ", task_group),
var = gsub("_", " ", var)) %>%
mutate(task_group = ifelse(task_group == "psychological refractory period two choices", "psychological refractory period", ifelse(task_group == "angling risk task always sunny", "angling risk task",task_group))) %>%
mutate(task_group = gsub("survey", "", task_group)) %>%
filter(task=="survey")%>%
arrange(task_group, rank)
labels = tmp %>%
distinct(dv, .keep_all=T)
p2 <- tmp %>%
ggplot(aes(x=factor(rank), y=value, fill=factor(key, levels = c("var_resid_pct", "var_ind_pct", "var_subs_pct"))))+
geom_bar(stat='identity', alpha = 0.75)+
geom_bar(stat='identity', color='#F8766D', show.legend=FALSE)+
scale_x_discrete(breaks = labels$rank,
labels = labels$var)+
coord_flip()+
facet_grid(task_group~., switch = "y", scales = "free_y", space = "free_y") +
theme(panel.spacing = unit(0.5, "lines"),
strip.placement = "outside",
strip.text.y = element_text(angle=180),
panel.background = element_rect(fill = NA),
panel.grid.major = element_line(colour = "grey85"),
legend.position = 'bottom')+
theme(legend.title = element_blank())+
scale_fill_manual(breaks = c("var_subs_pct", "var_ind_pct", "var_resid_pct"),
labels = c("Variance between individuals",
"Variance between sessions",
"Error variance"),
values=c("grey65", "grey45", "grey25"))+
ylab("")+
xlab("")
mylegend<-g_legend(p2)
p3 <- arrangeGrob(arrangeGrob(p1 +theme(legend.position="none"),
p2 + theme(legend.position="none"),
nrow=1),
mylegend, nrow=2,heights=c(10, 1))
ggsave(paste0('Variance_Breakdown_Plot.', out_device), plot = p3, device = out_device, path = fig_path, width = 24, height = 20, units = "in", dpi = img_dpi)
rm(tmp, labels, p1, p2 , p3)
| /code/figure_scripts/Variance_Breakdown_Plot.R | no_license | zenkavi/SRO_Retest_Analyses | R | false | false | 5,646 | r | source('/Users/zeynepenkavi/Dropbox/PoldrackLab/SRO_Retest_Analyses/code/figure_scripts/figure_res_wrapper.R')
if(!exists('rel_df')){
source('/Users/zeynepenkavi/Dropbox/PoldrackLab/SRO_Retest_Analyses/code/workspace_scripts/subject_data.R')
source('/Users/zeynepenkavi/Dropbox/PoldrackLab/SRO_Retest_Analyses/code/helper_functions/make_rel_df.R')
rel_df = make_rel_df(t1_df = test_data, t2_df = retest_data, metrics = c('spearman', 'icc', 'pearson', 'var_breakdown', 'partial_eta', 'sem'))
rel_df$task = 'task'
rel_df[grep('survey', rel_df$dv), 'task'] = 'survey'
rel_df[grep('holt', rel_df$dv), 'task'] = "task"
rel_df = rel_df %>%
select(dv, task, spearman, icc, pearson, partial_eta, sem, var_subs, var_ind, var_resid)
}
tmp = rel_df %>%
mutate(var_subs_pct = var_subs/(var_subs+var_ind+var_resid)*100,
var_ind_pct = var_ind/(var_subs+var_ind+var_resid)*100,
var_resid_pct = var_resid/(var_subs+var_ind+var_resid)*100) %>%
select(dv, task, var_subs_pct, var_ind_pct, var_resid_pct) %>%
mutate(dv = factor(dv, levels = dv[order(task)])) %>%
separate(dv, c("task_group", "var"), sep="\\.",remove=FALSE,extra="merge") %>%
mutate(task_group = factor(task_group, levels = task_group[order(task)])) %>%
arrange(task_group, var_subs_pct) %>%
mutate(rank = row_number()) %>%
arrange(task, task_group, rank) %>%
gather(key, value, -dv, -task_group, -var, -task, -rank) %>%
ungroup()%>%
mutate(task_group = gsub("_", " ", task_group),
var = gsub("_", " ", var)) %>%
mutate(task_group = ifelse(task_group == "psychological refractory period two choices", "psychological refractory period", ifelse(task_group == "angling risk task always sunny", "angling risk task",task_group))) %>%
mutate(task_group = gsub("survey", "", task_group)) %>%
filter(task=="task",
!grepl("EZ|hddm", dv))%>%
arrange(task_group, rank)
labels = tmp %>%
distinct(dv, .keep_all=T)
p1 <- tmp %>%
ggplot(aes(x=factor(rank), y=value, fill=factor(key, levels = c("var_resid_pct", "var_ind_pct", "var_subs_pct"))))+
geom_bar(stat='identity', alpha = 0.75, color='#00BFC4')+
scale_x_discrete(breaks = labels$rank,
labels = labels$var)+
coord_flip()+
facet_grid(task_group~., switch = "y", scales = "free_y", space = "free_y") +
theme(panel.spacing = unit(0.5, "lines"),
strip.placement = "outside",
strip.text.y = element_text(angle=180),
panel.background = element_rect(fill = NA),
panel.grid.major = element_line(colour = "grey85"),
legend.position = 'bottom')+
theme(legend.title = element_blank())+
scale_fill_manual(breaks = c("var_subs_pct", "var_ind_pct", "var_resid_pct"),
labels = c("Variance between individuals",
"Variance between sessions",
"Error variance"),
values=c("grey65", "grey45", "grey25"))+
ylab("")+
xlab("")
tmp = rel_df %>%
mutate(var_subs_pct = var_subs/(var_subs+var_ind+var_resid)*100,
var_ind_pct = var_ind/(var_subs+var_ind+var_resid)*100,
var_resid_pct = var_resid/(var_subs+var_ind+var_resid)*100) %>%
select(dv, task, var_subs_pct, var_ind_pct, var_resid_pct) %>%
mutate(dv = factor(dv, levels = dv[order(task)])) %>%
separate(dv, c("task_group", "var"), sep="\\.",remove=FALSE,extra="merge") %>%
mutate(task_group = factor(task_group, levels = task_group[order(task)])) %>%
arrange(task_group, var_subs_pct) %>%
mutate(rank = row_number()) %>%
arrange(task, task_group, rank) %>%
gather(key, value, -dv, -task_group, -var, -task, -rank) %>%
ungroup()%>%
mutate(task_group = gsub("_", " ", task_group),
var = gsub("_", " ", var)) %>%
mutate(task_group = ifelse(task_group == "psychological refractory period two choices", "psychological refractory period", ifelse(task_group == "angling risk task always sunny", "angling risk task",task_group))) %>%
mutate(task_group = gsub("survey", "", task_group)) %>%
filter(task=="survey")%>%
arrange(task_group, rank)
labels = tmp %>%
distinct(dv, .keep_all=T)
p2 <- tmp %>%
ggplot(aes(x=factor(rank), y=value, fill=factor(key, levels = c("var_resid_pct", "var_ind_pct", "var_subs_pct"))))+
geom_bar(stat='identity', alpha = 0.75)+
geom_bar(stat='identity', color='#F8766D', show.legend=FALSE)+
scale_x_discrete(breaks = labels$rank,
labels = labels$var)+
coord_flip()+
facet_grid(task_group~., switch = "y", scales = "free_y", space = "free_y") +
theme(panel.spacing = unit(0.5, "lines"),
strip.placement = "outside",
strip.text.y = element_text(angle=180),
panel.background = element_rect(fill = NA),
panel.grid.major = element_line(colour = "grey85"),
legend.position = 'bottom')+
theme(legend.title = element_blank())+
scale_fill_manual(breaks = c("var_subs_pct", "var_ind_pct", "var_resid_pct"),
labels = c("Variance between individuals",
"Variance between sessions",
"Error variance"),
values=c("grey65", "grey45", "grey25"))+
ylab("")+
xlab("")
mylegend<-g_legend(p2)
p3 <- arrangeGrob(arrangeGrob(p1 +theme(legend.position="none"),
p2 + theme(legend.position="none"),
nrow=1),
mylegend, nrow=2,heights=c(10, 1))
ggsave(paste0('Variance_Breakdown_Plot.', out_device), plot = p3, device = out_device, path = fig_path, width = 24, height = 20, units = "in", dpi = img_dpi)
rm(tmp, labels, p1, p2 , p3)
|
## Copyright (C) 2019 Prim'Act, Quentin Guibert
##
## This program is free software; you can redistribute it and/or modify it under
## the terms of the GNU General Public License as published by the Free Software
## Foundation; either version 3 of the License, or (at your option) any later
## version.
##
## This program is distributed in the hope that it will be useful, but WITHOUT
## ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
## FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
## details.
##
## You should have received a copy of the GNU General Public License along with
## this program; if not, see <https://www.gnu.org/licenses/>.
# Management rules: profit sharing
###--------------------------
# Initialisation
###--------------------------
# Get a object which contains the liability portfolio and related assumptions
central <- get(load(paste(racine@address$save_folder$central, "best_estimate.RData", sep = "/")))
ptf_passif <- central@canton@ptf_passif
# Current year
year <- 1L
###--------------------------
# Asset and liability before profit sharing
###--------------------------
# Liability portfolio at the end of the year before profit sharing
# ptf_eoy <- proj_annee_av_pb(an = year, x = ptf_passif, tx_soc = 0.155, coef_inf = 1, list_rd = c(0.02,0.01,0.01,0))
liab_eoy <- viellissement_av_pb(an = year, ptf_passif, coef_inf = 1, list_rd =c(0.02,0.01,0.01,0), tx_soc = 0.155)
asset_eoy <- update_PortFin(an = year, ptf_fin, new_mp_ESG = central@canton@mp_esg,
flux_milieu = liab_eoy[["flux_milieu"]], flux_fin = liab_eoy[["flux_fin"]])
# Get the updated asset portfolio
ptf_fin <- asset_eoy[["ptf"]]
# Get the financial incomes and the realized capital gain and loss on bonds
income <- asset_eoy[["revenu_fin"]]
var_vnc_bond <- asset_eoy[["var_vnc_oblig"]]
# Reallocate the asset portfolio
asset_realloc <- reallocate(ptf_fin, central@canton@param_alm@ptf_reference, central@canton@param_alm@alloc_cible)
# Compute the technical results, including Delta provisions on 'PRE'
result_tech <- calc_result_technique(liab_eoy, asset_realloc[["var_pre"]])
print(resultat_tech)
# Compute the financial results
result_fin <- calc_resultat_fin(income + var_vnc_bond, asset_realloc[["pmvr"]],
frais_fin = 0, asset_realloc[["var_rc"]])
# Compute the TRA (return rate on assets according to the Frenc GAAP)
tra <- calc_tra(asset_realloc[["plac_moy_vnc"]], result_fin)
print(tra)
###--------------------------
# Apply the profit sharing algorithm
###--------------------------
result_revalo <- calc_revalo(central@canton, liab_eoy, tra,
asset_realloc[["plac_moy_vnc"]], result_tech)
#updated PPB
result_revalo$ppb
# Profit sharing rate
result_revalo$tx_pb
# Amont of profil sharing to allocate
result_revalo$add_rev_nette_stock
# Amount of gain of loss which are be realized for reaching the target revalorisation rate
result_revalo$pmvl_liq
###--------------------------
# Explore the profit sharing algorithm with simple examples
###--------------------------
# Step 0: initialisation
###-------------------------
# Data
tra_1 <- 0.05
tra_2 <- 0.02
tra_3 <- 0.01
tra_4 <- - 0.01
# 4 products and their profit sharing rates
pm_moy <- rep(100, 4)
tx_pb <- c(0.90, 0.95, 0.97, 1)
# Create an initial PPB as nul
ppb <- new(Class = "Ppb")
ppb@valeur_ppb <- ppb@ppb_debut <- 8 # The initial amount of PPB is equal to 8
ppb@hist_ppb <- rep(1, 8) # This amount have be endowed uniformaly during the last 8 years
ppb@seuil_rep <- ppb@seuil_dot <- 0.5 # The PPB can be endowed or ceded until 50%
# Assume the loadings are nul
tx_enc_moy <- c(0, 0, 0, 0)
# Step 1: contractual profit sharing
###-------------------------
# Compute the financial result related to the liability
# base_fin_1 <- tra_1 * (sum(pm_moy) + ppb["ppb_debut"]) * pm_moy / sum(pm_moy)
base_fin_1 <- base_prod_fin(tra_1, pm_moy, ppb)
base_fin_2 <- base_prod_fin(tra_2, pm_moy, ppb)
base_fin_3 <- base_prod_fin(tra_3, pm_moy, ppb)
base_fin_4 <- base_prod_fin(tra_4, pm_moy, ppb)
# Revalorisation considering a minimum rate of 1%
rev_stock_brut <- pm_moy * 0.01
ch_enc_th <- pm_moy * (1 + 0.01) * tx_enc_moy
# Amount of contractual profit sharing considering the minimal rate
reval_contr_1 <- pb_contr(base_fin_1$base_prod_fin, tx_pb, rev_stock_brut, ch_enc_th, tx_enc_moy)
reval_contr_2 <- pb_contr(base_fin_2$base_prod_fin, tx_pb, rev_stock_brut, ch_enc_th, tx_enc_moy)
reval_contr_3 <- pb_contr(base_fin_3$base_prod_fin, tx_pb, rev_stock_brut, ch_enc_th, tx_enc_moy)
reval_contr_4 <- pb_contr(base_fin_4$base_prod_fin, tx_pb, rev_stock_brut, ch_enc_th, tx_enc_moy)
# Step 2: TMG
###-------------------------
# The PPB can finance the need ralated to the TMG garantee
tmg <- 0.02 # High TMG
# tmg <- 0 # No TMG
bes_tmg_stock <- pm_moy * tmg
bes_tmg_prest <- pm_moy * 0 # Assumption no TMG on benefits
financement_tmg <- finance_tmg(bes_tmg_prest, bes_tmg_stock, ppb)
# Update the PPB objet
ppb <- financement_tmg[["ppb"]]
# Step 3: Regulatory constraints on the PPB
###-------------------------
# Regulatory constraint: 8 years for using the PPB
ppb_8 <- ppb_8ans(ppb)
# Update the PPB objet
ppb <- ppb_8[["ppb"]]
# Allocate this amount to each product, e.g. with the weights in terms of MP
som <- sum(pm_moy)
ppb8_ind <- ppb_8$ppb_8 * pm_moy / som
# Step 4: Reach the target rate
###-------------------------
# Assume the insured expects a rate on return of 3.00%
target_rate <- 0.03
bes_tx_cible <- pm_moy * target_rate
# Option #1 Use the PPB for reaching this target
tx_cibl_ppb_1 <- finance_cible_ppb(bes_tx_cible, reval_contr_1$rev_stock_nette_contr, ppb, ppb8_ind)
tx_cibl_ppb_2 <- finance_cible_ppb(bes_tx_cible, reval_contr_2$rev_stock_nette_contr, ppb, ppb8_ind)
tx_cibl_ppb_3 <- finance_cible_ppb(bes_tx_cible, reval_contr_3$rev_stock_nette_contr, ppb, ppb8_ind)
tx_cibl_ppb_4 <- finance_cible_ppb(bes_tx_cible, reval_contr_4$rev_stock_nette_contr, ppb, ppb8_ind)
# Mise a jour de la PPB, e.g. in case #1
ppb <- tx_cibl_ppb_1$ppb
# Option #2 Sell shares for reaching this target
# Assume the amount of unrealized gain of shares which can be sold is limited to 5
seuil_pmvl <- 5
# Revalorisation after selling shares
tx_cibl_pmvl_1 <- finance_cible_pmvl(bes_tx_cible, tx_cibl_ppb_1$rev_stock_nette, base_fin_1$base_prod_fin, seuil_pmvl, tx_pb)
tx_cibl_pmvl_2 <- finance_cible_pmvl(bes_tx_cible, tx_cibl_ppb_2$rev_stock_nette, base_fin_2$base_prod_fin, seuil_pmvl, tx_pb)
tx_cibl_pmvl_3 <- finance_cible_pmvl(bes_tx_cible, tx_cibl_ppb_3$rev_stock_nette, base_fin_3$base_prod_fin, seuil_pmvl, tx_pb)
tx_cibl_pmvl_4 <- finance_cible_pmvl(bes_tx_cible, tx_cibl_ppb_4$rev_stock_nette, base_fin_4$base_prod_fin, seuil_pmvl, tx_pb)
# Step 5: Apply the legal constraints on the overall portfolio
#---------------------------------------------------------------
# Technical results is zero
result_tech <- 0
it_tech <- rev_stock_brut
revalo_finale_1 <- finance_contrainte_legale(base_fin_1$base_prod_fin, base_fin_1$base_prod_fin_port,
result_tech, it_tech,
tx_cibl_pmvl_1$rev_stock_nette,
bes_tmg_prest,
tx_cibl_ppb_1$dotation, 0, ppb,
central@canton@param_revalo)
###--------------------------
# Summary function for launching all the steps over the year
###--------------------------
result_proj_an <-proj_an(x = central@canton, annee_fin = central@param_be@nb_annee, pre_on = T)
| /5-profit_sharing.R | no_license | RebeccaWel/Environnement | R | false | false | 7,869 | r | ## Copyright (C) 2019 Prim'Act, Quentin Guibert
##
## This program is free software; you can redistribute it and/or modify it under
## the terms of the GNU General Public License as published by the Free Software
## Foundation; either version 3 of the License, or (at your option) any later
## version.
##
## This program is distributed in the hope that it will be useful, but WITHOUT
## ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
## FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
## details.
##
## You should have received a copy of the GNU General Public License along with
## this program; if not, see <https://www.gnu.org/licenses/>.
# Management rules: profit sharing
###--------------------------
# Initialisation
###--------------------------
# Get a object which contains the liability portfolio and related assumptions
central <- get(load(paste(racine@address$save_folder$central, "best_estimate.RData", sep = "/")))
ptf_passif <- central@canton@ptf_passif
# Current year
year <- 1L
###--------------------------
# Asset and liability before profit sharing
###--------------------------
# Liability portfolio at the end of the year before profit sharing
# ptf_eoy <- proj_annee_av_pb(an = year, x = ptf_passif, tx_soc = 0.155, coef_inf = 1, list_rd = c(0.02,0.01,0.01,0))
liab_eoy <- viellissement_av_pb(an = year, ptf_passif, coef_inf = 1, list_rd =c(0.02,0.01,0.01,0), tx_soc = 0.155)
asset_eoy <- update_PortFin(an = year, ptf_fin, new_mp_ESG = central@canton@mp_esg,
flux_milieu = liab_eoy[["flux_milieu"]], flux_fin = liab_eoy[["flux_fin"]])
# Get the updated asset portfolio
ptf_fin <- asset_eoy[["ptf"]]
# Get the financial incomes and the realized capital gain and loss on bonds
income <- asset_eoy[["revenu_fin"]]
var_vnc_bond <- asset_eoy[["var_vnc_oblig"]]
# Reallocate the asset portfolio
asset_realloc <- reallocate(ptf_fin, central@canton@param_alm@ptf_reference, central@canton@param_alm@alloc_cible)
# Compute the technical results, including Delta provisions on 'PRE'
result_tech <- calc_result_technique(liab_eoy, asset_realloc[["var_pre"]])
print(resultat_tech)
# Compute the financial results
result_fin <- calc_resultat_fin(income + var_vnc_bond, asset_realloc[["pmvr"]],
frais_fin = 0, asset_realloc[["var_rc"]])
# Compute the TRA (return rate on assets according to the Frenc GAAP)
tra <- calc_tra(asset_realloc[["plac_moy_vnc"]], result_fin)
print(tra)
###--------------------------
# Apply the profit sharing algorithm
###--------------------------
result_revalo <- calc_revalo(central@canton, liab_eoy, tra,
asset_realloc[["plac_moy_vnc"]], result_tech)
#updated PPB
result_revalo$ppb
# Profit sharing rate
result_revalo$tx_pb
# Amont of profil sharing to allocate
result_revalo$add_rev_nette_stock
# Amount of gain of loss which are be realized for reaching the target revalorisation rate
result_revalo$pmvl_liq
###--------------------------
# Explore the profit sharing algorithm with simple examples
###--------------------------
# Step 0: initialisation
###-------------------------
# Data
tra_1 <- 0.05
tra_2 <- 0.02
tra_3 <- 0.01
tra_4 <- - 0.01
# 4 products and their profit sharing rates
pm_moy <- rep(100, 4)
tx_pb <- c(0.90, 0.95, 0.97, 1)
# Create an initial PPB as nul
ppb <- new(Class = "Ppb")
ppb@valeur_ppb <- ppb@ppb_debut <- 8 # The initial amount of PPB is equal to 8
ppb@hist_ppb <- rep(1, 8) # This amount have be endowed uniformaly during the last 8 years
ppb@seuil_rep <- ppb@seuil_dot <- 0.5 # The PPB can be endowed or ceded until 50%
# Assume the loadings are nul
tx_enc_moy <- c(0, 0, 0, 0)
# Step 1: contractual profit sharing
###-------------------------
# Compute the financial result related to the liability
# base_fin_1 <- tra_1 * (sum(pm_moy) + ppb["ppb_debut"]) * pm_moy / sum(pm_moy)
base_fin_1 <- base_prod_fin(tra_1, pm_moy, ppb)
base_fin_2 <- base_prod_fin(tra_2, pm_moy, ppb)
base_fin_3 <- base_prod_fin(tra_3, pm_moy, ppb)
base_fin_4 <- base_prod_fin(tra_4, pm_moy, ppb)
# Revalorisation considering a minimum rate of 1%
rev_stock_brut <- pm_moy * 0.01
ch_enc_th <- pm_moy * (1 + 0.01) * tx_enc_moy
# Amount of contractual profit sharing considering the minimal rate
reval_contr_1 <- pb_contr(base_fin_1$base_prod_fin, tx_pb, rev_stock_brut, ch_enc_th, tx_enc_moy)
reval_contr_2 <- pb_contr(base_fin_2$base_prod_fin, tx_pb, rev_stock_brut, ch_enc_th, tx_enc_moy)
reval_contr_3 <- pb_contr(base_fin_3$base_prod_fin, tx_pb, rev_stock_brut, ch_enc_th, tx_enc_moy)
reval_contr_4 <- pb_contr(base_fin_4$base_prod_fin, tx_pb, rev_stock_brut, ch_enc_th, tx_enc_moy)
# Step 2: TMG
###-------------------------
# The PPB can finance the need ralated to the TMG garantee
tmg <- 0.02 # High TMG
# tmg <- 0 # No TMG
bes_tmg_stock <- pm_moy * tmg
bes_tmg_prest <- pm_moy * 0 # Assumption no TMG on benefits
financement_tmg <- finance_tmg(bes_tmg_prest, bes_tmg_stock, ppb)
# Update the PPB objet
ppb <- financement_tmg[["ppb"]]
# Step 3: Regulatory constraints on the PPB
###-------------------------
# Regulatory constraint: 8 years for using the PPB
ppb_8 <- ppb_8ans(ppb)
# Update the PPB objet
ppb <- ppb_8[["ppb"]]
# Allocate this amount to each product, e.g. with the weights in terms of MP
som <- sum(pm_moy)
ppb8_ind <- ppb_8$ppb_8 * pm_moy / som
# Step 4: Reach the target rate
###-------------------------
# Assume the insured expects a rate on return of 3.00%
target_rate <- 0.03
bes_tx_cible <- pm_moy * target_rate
# Option #1 Use the PPB for reaching this target
tx_cibl_ppb_1 <- finance_cible_ppb(bes_tx_cible, reval_contr_1$rev_stock_nette_contr, ppb, ppb8_ind)
tx_cibl_ppb_2 <- finance_cible_ppb(bes_tx_cible, reval_contr_2$rev_stock_nette_contr, ppb, ppb8_ind)
tx_cibl_ppb_3 <- finance_cible_ppb(bes_tx_cible, reval_contr_3$rev_stock_nette_contr, ppb, ppb8_ind)
tx_cibl_ppb_4 <- finance_cible_ppb(bes_tx_cible, reval_contr_4$rev_stock_nette_contr, ppb, ppb8_ind)
# Mise a jour de la PPB, e.g. in case #1
ppb <- tx_cibl_ppb_1$ppb
# Option #2 Sell shares for reaching this target
# Assume the amount of unrealized gain of shares which can be sold is limited to 5
seuil_pmvl <- 5
# Revalorisation after selling shares
tx_cibl_pmvl_1 <- finance_cible_pmvl(bes_tx_cible, tx_cibl_ppb_1$rev_stock_nette, base_fin_1$base_prod_fin, seuil_pmvl, tx_pb)
tx_cibl_pmvl_2 <- finance_cible_pmvl(bes_tx_cible, tx_cibl_ppb_2$rev_stock_nette, base_fin_2$base_prod_fin, seuil_pmvl, tx_pb)
tx_cibl_pmvl_3 <- finance_cible_pmvl(bes_tx_cible, tx_cibl_ppb_3$rev_stock_nette, base_fin_3$base_prod_fin, seuil_pmvl, tx_pb)
tx_cibl_pmvl_4 <- finance_cible_pmvl(bes_tx_cible, tx_cibl_ppb_4$rev_stock_nette, base_fin_4$base_prod_fin, seuil_pmvl, tx_pb)
# Step 5: Apply the legal constraints on the overall portfolio
#---------------------------------------------------------------
# Technical results is zero
result_tech <- 0
it_tech <- rev_stock_brut
revalo_finale_1 <- finance_contrainte_legale(base_fin_1$base_prod_fin, base_fin_1$base_prod_fin_port,
result_tech, it_tech,
tx_cibl_pmvl_1$rev_stock_nette,
bes_tmg_prest,
tx_cibl_ppb_1$dotation, 0, ppb,
central@canton@param_revalo)
###--------------------------
# Summary function for launching all the steps over the year
###--------------------------
result_proj_an <-proj_an(x = central@canton, annee_fin = central@param_be@nb_annee, pre_on = T)
|
xdistance <- function(x, y, method = "euclidean") {
# calculate cross-dissimilarities between
# rows of x and rows of y
# returns a nonsymmetric matrix where
# d[a, b] is the dissimilarity between
# x[a, ] and y[b, ]
# Sarah Goslee 2017-02-17, modified from legacy Splus code dated 01/01/01
if(is.null(ncol(x))) {
x <- matrix(x, ncol=1)
rownames(x) <- seq_len(nrow(x))
}
if(is.null(ncol(y))) {
y <- matrix(y, ncol=1)
rownames(y) <- seq_len(nrow(y))
}
if(!(ncol(x) == ncol(y)))
stop("Matrices must have the same number of columns\n")
x.names <- paste0("x", row.names(x))
y.names <- paste0("y", row.names(y))
x <- as.matrix(x)
y <- as.matrix(y)
d <- rbind(x, y)
d <- full(distance(d, method=method))
d <- d[seq(1, nrow(x)), seq(nrow(x) + 1, nrow(x) + nrow(y)), drop=FALSE]
rownames(d) <- x.names
colnames(d) <- y.names
class(d) <- "xdist"
d
}
| /R/xdistance.R | no_license | cran/ecodist | R | false | false | 980 | r | xdistance <- function(x, y, method = "euclidean") {
# calculate cross-dissimilarities between
# rows of x and rows of y
# returns a nonsymmetric matrix where
# d[a, b] is the dissimilarity between
# x[a, ] and y[b, ]
# Sarah Goslee 2017-02-17, modified from legacy Splus code dated 01/01/01
if(is.null(ncol(x))) {
x <- matrix(x, ncol=1)
rownames(x) <- seq_len(nrow(x))
}
if(is.null(ncol(y))) {
y <- matrix(y, ncol=1)
rownames(y) <- seq_len(nrow(y))
}
if(!(ncol(x) == ncol(y)))
stop("Matrices must have the same number of columns\n")
x.names <- paste0("x", row.names(x))
y.names <- paste0("y", row.names(y))
x <- as.matrix(x)
y <- as.matrix(y)
d <- rbind(x, y)
d <- full(distance(d, method=method))
d <- d[seq(1, nrow(x)), seq(nrow(x) + 1, nrow(x) + nrow(y)), drop=FALSE]
rownames(d) <- x.names
colnames(d) <- y.names
class(d) <- "xdist"
d
}
|
# Copyright 2019 Battelle Memorial Institute; see the LICENSE file.
#' module_water_water_demand_industry_xml
#'
#' Construct XML data structure for \code{water_demand_industry.xml}.
#'
#' @param command API command to execute
#' @param ... other optional parameters, depending on command
#' @return Depends on \code{command}: either a vector of required inputs,
#' a vector of output names, or (if \code{command} is "MAKE") all
#' the generated outputs: \code{water_demand_industry.xml}. The corresponding file in the
#' original data system was \code{batch_water_demand_industry.xml.R} (water XML).
module_water_water_demand_industry_xml <- function(command, ...) {
if(command == driver.DECLARE_INPUTS) {
return(c("L232.TechCoef"))
} else if(command == driver.DECLARE_OUTPUTS) {
return(c(XML = "water_demand_industry.xml"))
} else if(command == driver.MAKE) {
all_data <- list(...)[[1]]
# Load required inputs
L232.TechCoef <- get_data(all_data, "L232.TechCoef")
# ===================================================
# Produce outputs
create_xml("water_demand_industry.xml") %>%
add_xml_data(L232.TechCoef, "TechCoef") %>%
add_precursors("L232.TechCoef") ->
water_demand_industry.xml
return_data(water_demand_industry.xml)
} else {
stop("Unknown command")
}
}
| /input/gcamdata/R/zwater_xml_water_demand_industry.R | permissive | JGCRI/gcam-core | R | false | false | 1,336 | r | # Copyright 2019 Battelle Memorial Institute; see the LICENSE file.
#' module_water_water_demand_industry_xml
#'
#' Construct XML data structure for \code{water_demand_industry.xml}.
#'
#' @param command API command to execute
#' @param ... other optional parameters, depending on command
#' @return Depends on \code{command}: either a vector of required inputs,
#' a vector of output names, or (if \code{command} is "MAKE") all
#' the generated outputs: \code{water_demand_industry.xml}. The corresponding file in the
#' original data system was \code{batch_water_demand_industry.xml.R} (water XML).
module_water_water_demand_industry_xml <- function(command, ...) {
if(command == driver.DECLARE_INPUTS) {
return(c("L232.TechCoef"))
} else if(command == driver.DECLARE_OUTPUTS) {
return(c(XML = "water_demand_industry.xml"))
} else if(command == driver.MAKE) {
all_data <- list(...)[[1]]
# Load required inputs
L232.TechCoef <- get_data(all_data, "L232.TechCoef")
# ===================================================
# Produce outputs
create_xml("water_demand_industry.xml") %>%
add_xml_data(L232.TechCoef, "TechCoef") %>%
add_precursors("L232.TechCoef") ->
water_demand_industry.xml
return_data(water_demand_industry.xml)
} else {
stop("Unknown command")
}
}
|
library("ggspatial")
library("ggplot2")
theme_set(theme_bw())
library("sf")
library("rnaturalearth")
library("rnaturalearthdata")
world <- ne_countries(scale = "medium", returnclass = "sf")
class(world)
# black and white map with compass and distance legend
ggplot(data = world) +
geom_sf() +
annotation_scale(location = "bl", width_hint = 0.5) +
annotation_north_arrow(location = "bl", which_north = "true",
pad_x = unit(0.75, "in"), pad_y = unit(0.5, "in"),
style = north_arrow_fancy_orienteering) +
coord_sf(xlim = c(-102.15, -74.12), ylim = c(7.65, 33.97))
## Scale on map varies by more than 10%, scale bar may be inaccurate
# with country names
world_points<- st_centroid(world)
world_points <- cbind(world, st_coordinates(st_centroid(world$geometry)))
ggplot(data = world) +
geom_sf() +
geom_text(data= world_points,aes(x=X, y=Y, label=name),
color = "darkblue", fontface = "bold", check_overlap = FALSE) +
annotate(geom = "text", x = -90, y = 26, label = "Gulf of Mexico",
fontface = "italic", color = "grey22", size = 6) +
coord_sf(xlim = c(-102.15, -74.12), ylim = c(7.65, 33.97), expand = FALSE)
ggplot(data = world) +
geom_sf(fill= 'antiquewhite') +
geom_text(data= world_points,aes(x=X, y=Y, label=name),
color = 'darkblue', fontface = 'bold',
check_overlap = FALSE) + annotate(geom = 'text',
x = -300, y = 400, label = 'Gulf of Mexico',
fontface = 'italic', color = 'grey22', size = 6) +
annotation_scale(location = 'bl', width_hint = 0.5) +
annotation_north_arrow(location = 'bl',
which_north = 'true', pad_x = unit(0.75, 'in'),
pad_y = unit(0.5, 'in'), style = north_arrow_fancy_orienteering) +
coord_sf(xlim = c(-102.15, -74.12),
ylim = c(7.65, 33.97), expand = FALSE) +
xlab('Longitude') + ylab('Latitude') +
ggtitle('Map of the Gulf of Mexico and the Caribbean Sea') +
theme(panel.grid.major = element_line(color = gray(.5), linetype = 'dashed',
size = 0.5), panel.background = element_rect(fill = 'aliceblue'))
| /charts/map.with.compass.and.distance.r | no_license | gsdavis1959/R_examples | R | false | false | 2,074 | r | library("ggspatial")
library("ggplot2")
theme_set(theme_bw())
library("sf")
library("rnaturalearth")
library("rnaturalearthdata")
world <- ne_countries(scale = "medium", returnclass = "sf")
class(world)
# black and white map with compass and distance legend
ggplot(data = world) +
geom_sf() +
annotation_scale(location = "bl", width_hint = 0.5) +
annotation_north_arrow(location = "bl", which_north = "true",
pad_x = unit(0.75, "in"), pad_y = unit(0.5, "in"),
style = north_arrow_fancy_orienteering) +
coord_sf(xlim = c(-102.15, -74.12), ylim = c(7.65, 33.97))
## Scale on map varies by more than 10%, scale bar may be inaccurate
# with country names
world_points<- st_centroid(world)
world_points <- cbind(world, st_coordinates(st_centroid(world$geometry)))
ggplot(data = world) +
geom_sf() +
geom_text(data= world_points,aes(x=X, y=Y, label=name),
color = "darkblue", fontface = "bold", check_overlap = FALSE) +
annotate(geom = "text", x = -90, y = 26, label = "Gulf of Mexico",
fontface = "italic", color = "grey22", size = 6) +
coord_sf(xlim = c(-102.15, -74.12), ylim = c(7.65, 33.97), expand = FALSE)
ggplot(data = world) +
geom_sf(fill= 'antiquewhite') +
geom_text(data= world_points,aes(x=X, y=Y, label=name),
color = 'darkblue', fontface = 'bold',
check_overlap = FALSE) + annotate(geom = 'text',
x = -300, y = 400, label = 'Gulf of Mexico',
fontface = 'italic', color = 'grey22', size = 6) +
annotation_scale(location = 'bl', width_hint = 0.5) +
annotation_north_arrow(location = 'bl',
which_north = 'true', pad_x = unit(0.75, 'in'),
pad_y = unit(0.5, 'in'), style = north_arrow_fancy_orienteering) +
coord_sf(xlim = c(-102.15, -74.12),
ylim = c(7.65, 33.97), expand = FALSE) +
xlab('Longitude') + ylab('Latitude') +
ggtitle('Map of the Gulf of Mexico and the Caribbean Sea') +
theme(panel.grid.major = element_line(color = gray(.5), linetype = 'dashed',
size = 0.5), panel.background = element_rect(fill = 'aliceblue'))
|
plot4 <- function(ucifile = "C:/Users/cwilhelm/Documents/Coursera/exdata%2Fdata%2Fhousehold_power_consumption/household_power_consumption.txt") {
ucidata <- read.table(ucifile,header=TRUE,sep=";",na.strings="?")
subfirst <- subset(ucidata,Date=="1/2/2007")
subsecond <- subset(ucidata,Date=="2/2/2007")
subdata <- rbind(subfirst,subsecond)
subdata$Timestamp <- paste(subdata$Date, subdata$Time)
png("plot4.png",height=480,width=480)
par(mfrow=c(2,2))
plot(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Global_active_power,type="l",xlab="",ylab="Global Active Power")
plot(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Voltage,type="l",xlab="datetime",ylab="Voltage")
plot(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Sub_metering_1,type="l",xlab="",ylab="Energy sub metering")
points(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Sub_metering_2,type="l",col="red")
points(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Sub_metering_3,type="l",col="blue")
legend("topright",legend=c("Sub_metering_1","Sub_metering_2","Sub_metering_3"),col=c("black","red","blue"),lty=1,bty="n")
plot(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Global_reactive_power,type="l",xlab="datetime",ylab="Global_reactive_power")
dev.off()
} | /Plot4.R | no_license | cwilhelm486/ExData_Plotting1 | R | false | false | 1,324 | r | plot4 <- function(ucifile = "C:/Users/cwilhelm/Documents/Coursera/exdata%2Fdata%2Fhousehold_power_consumption/household_power_consumption.txt") {
ucidata <- read.table(ucifile,header=TRUE,sep=";",na.strings="?")
subfirst <- subset(ucidata,Date=="1/2/2007")
subsecond <- subset(ucidata,Date=="2/2/2007")
subdata <- rbind(subfirst,subsecond)
subdata$Timestamp <- paste(subdata$Date, subdata$Time)
png("plot4.png",height=480,width=480)
par(mfrow=c(2,2))
plot(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Global_active_power,type="l",xlab="",ylab="Global Active Power")
plot(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Voltage,type="l",xlab="datetime",ylab="Voltage")
plot(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Sub_metering_1,type="l",xlab="",ylab="Energy sub metering")
points(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Sub_metering_2,type="l",col="red")
points(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Sub_metering_3,type="l",col="blue")
legend("topright",legend=c("Sub_metering_1","Sub_metering_2","Sub_metering_3"),col=c("black","red","blue"),lty=1,bty="n")
plot(strptime(subdata$Timestamp, "%d/%m/%Y %H:%M:%S"),subdata$Global_reactive_power,type="l",xlab="datetime",ylab="Global_reactive_power")
dev.off()
} |
\name{FastRCS-package}
\alias{FastPCS-package}
\docType{package}
\title{Code to compute the FastRCS regression outlyingness index.}
\description{
Uses the FastRCS algorithm to compute the RCS outlyingness index
of regression.
}
\details{
\tabular{ll}{
Package: \tab FastRCS\cr
Type: \tab Package\cr
Version: \tab 0.1.1\cr
Date: \tab 2013-01-13\cr
Suggests: \tab mvtnorm\cr
License: \tab GPL (>= 2)\cr
LazyLoad: \tab yes\cr
}
Index:
\preformatted{
FastRCS Function to compute the FastRCS regression outlyingness
index.
FRCSnumStarts Internal function used to compute the FastRCS regression
outlyingness index.
plot.FastRCS Robust Diagnostic Plots For FastRCS.
quanf Internal function used to compute the FastRCS regression
outlyingness index.
}
}
\references{
Vakili, K. and Schmitt, E. (2014).
Finding Regression Outliers With FastRCS.
(http://arxiv.org/abs/1307.4834)
}
\author{
Kaveh Vakili [aut, cre],
Maintainer: Kaveh Vakili <vakili.kaveh.email@gmail.com>
}
\keyword{package}
| /fuzzedpackages/FastRCS/man/FastRCS-package.Rd | no_license | akhikolla/testpackages | R | false | false | 1,109 | rd | \name{FastRCS-package}
\alias{FastPCS-package}
\docType{package}
\title{Code to compute the FastRCS regression outlyingness index.}
\description{
Uses the FastRCS algorithm to compute the RCS outlyingness index
of regression.
}
\details{
\tabular{ll}{
Package: \tab FastRCS\cr
Type: \tab Package\cr
Version: \tab 0.1.1\cr
Date: \tab 2013-01-13\cr
Suggests: \tab mvtnorm\cr
License: \tab GPL (>= 2)\cr
LazyLoad: \tab yes\cr
}
Index:
\preformatted{
FastRCS Function to compute the FastRCS regression outlyingness
index.
FRCSnumStarts Internal function used to compute the FastRCS regression
outlyingness index.
plot.FastRCS Robust Diagnostic Plots For FastRCS.
quanf Internal function used to compute the FastRCS regression
outlyingness index.
}
}
\references{
Vakili, K. and Schmitt, E. (2014).
Finding Regression Outliers With FastRCS.
(http://arxiv.org/abs/1307.4834)
}
\author{
Kaveh Vakili [aut, cre],
Maintainer: Kaveh Vakili <vakili.kaveh.email@gmail.com>
}
\keyword{package}
|
#' @title To explain how to initialize clusters for the divisive algorithm.
#' @description To explain how to initialize clusters for the divisive algorithm.
#' @param initList is a clusters list. It will contain clusters with one element.
#' @details This function will explain how to calculate every cluster that can be created by joining initial clusters with each other. It creates clusters
#' from length = 1 until a cluster with every element is created.
#' @details These clusters will be used to find the most different clusters that we can create by dividing the initial cluster.
#' @author Roberto AlcΓ‘ntara \email{roberto.alcantara@@edu.uah.es}
#' @author Juan JosΓ© Cuadrado \email{jjcg@@uah.es}
#' @author Universidad de AlcalΓ‘ de Henares
#' @return A cluster list. Explanation.
#' @examples
#'
#' data <- c(1:8)
#'
#' matrix <- matrix(data, ncol=2)
#'
#' listData <- toListDivisive(data)
#'
#' listMatrix <- toListDivisive(matrix)
#'
#' initClusters.details(listData)
#'
#' initClusters.details(listMatrix)
#'
#' @export
initClusters.details <- function(initList){
message("\n 'initClusters' method initializes the clusters used in the divisive algorithm. \n\n")
message("\n To know which are the most different clusters, we need to know the distance between \n every possible clusters that could be created with the initial elements. \n")
message("\n This step is the most computationally complex, so it will make the algorithm to get the \n solution with delay, or even, not to find a solution because of the computers capacities.\n\n")
clusters <- initList
unitClusters <- initList
aux <- initList
goal <- initList[[length(initList)]]
res <- c()
auxAux <- c()
while(nrow(clusters[[length(clusters)]]) != length(unitClusters)){
for(i in seq_len(length(aux))){
lastElement <- aux[[i]]
lastElement <- (lastElement[nrow(lastElement),])
lastElement <- matrix(lastElement,ncol=2)
clusterIndex <- getClusterIndex(unitClusters,lastElement)
for (j in 1:length(unitClusters)){
cluster1 <- aux[[i]]
cluster1 <- matrix(cluster1,ncol=2)
cluster2 <- unitClusters[[j]]
if (j > clusterIndex){
newCluster <- c()
for (k in (1:nrow(cluster1))){
newCluster <- c(newCluster,cluster1[k,])
}
for (k in (1:nrow(cluster2))){
newCluster <- c(newCluster,cluster2[k,])
}
newCluster <- matrix(newCluster, ncol=2, byrow=TRUE)
res[[length(res) + 1]] <- newCluster
data <- newCluster[nrow(newCluster),]
asMatrix <- matrix(data,ncol=2)
if(!equalCluster(asMatrix,goal)){
auxAux[[length(auxAux) + 1]] <- newCluster
}
}
}
}
aux <- auxAux
auxAux <- c()
clusters <- c(clusters, res)
res <- c()
}
message("\n The clusters created using \n")
print(unitClusters)
message("\n are: \n")
print(clusters)
clusters
}
| /R/initClusters.details.R | no_license | cran/LearnClust | R | false | false | 3,054 | r | #' @title To explain how to initialize clusters for the divisive algorithm.
#' @description To explain how to initialize clusters for the divisive algorithm.
#' @param initList is a clusters list. It will contain clusters with one element.
#' @details This function will explain how to calculate every cluster that can be created by joining initial clusters with each other. It creates clusters
#' from length = 1 until a cluster with every element is created.
#' @details These clusters will be used to find the most different clusters that we can create by dividing the initial cluster.
#' @author Roberto AlcΓ‘ntara \email{roberto.alcantara@@edu.uah.es}
#' @author Juan JosΓ© Cuadrado \email{jjcg@@uah.es}
#' @author Universidad de AlcalΓ‘ de Henares
#' @return A cluster list. Explanation.
#' @examples
#'
#' data <- c(1:8)
#'
#' matrix <- matrix(data, ncol=2)
#'
#' listData <- toListDivisive(data)
#'
#' listMatrix <- toListDivisive(matrix)
#'
#' initClusters.details(listData)
#'
#' initClusters.details(listMatrix)
#'
#' @export
initClusters.details <- function(initList){
message("\n 'initClusters' method initializes the clusters used in the divisive algorithm. \n\n")
message("\n To know which are the most different clusters, we need to know the distance between \n every possible clusters that could be created with the initial elements. \n")
message("\n This step is the most computationally complex, so it will make the algorithm to get the \n solution with delay, or even, not to find a solution because of the computers capacities.\n\n")
clusters <- initList
unitClusters <- initList
aux <- initList
goal <- initList[[length(initList)]]
res <- c()
auxAux <- c()
while(nrow(clusters[[length(clusters)]]) != length(unitClusters)){
for(i in seq_len(length(aux))){
lastElement <- aux[[i]]
lastElement <- (lastElement[nrow(lastElement),])
lastElement <- matrix(lastElement,ncol=2)
clusterIndex <- getClusterIndex(unitClusters,lastElement)
for (j in 1:length(unitClusters)){
cluster1 <- aux[[i]]
cluster1 <- matrix(cluster1,ncol=2)
cluster2 <- unitClusters[[j]]
if (j > clusterIndex){
newCluster <- c()
for (k in (1:nrow(cluster1))){
newCluster <- c(newCluster,cluster1[k,])
}
for (k in (1:nrow(cluster2))){
newCluster <- c(newCluster,cluster2[k,])
}
newCluster <- matrix(newCluster, ncol=2, byrow=TRUE)
res[[length(res) + 1]] <- newCluster
data <- newCluster[nrow(newCluster),]
asMatrix <- matrix(data,ncol=2)
if(!equalCluster(asMatrix,goal)){
auxAux[[length(auxAux) + 1]] <- newCluster
}
}
}
}
aux <- auxAux
auxAux <- c()
clusters <- c(clusters, res)
res <- c()
}
message("\n The clusters created using \n")
print(unitClusters)
message("\n are: \n")
print(clusters)
clusters
}
|
% Generated by roxygen2 (4.1.0): do not edit by hand
% Please edit documentation in R/fsldilate.R
\name{fsldilate}
\alias{fsldilate}
\title{Dilate image using FSL}
\usage{
fsldilate(file, outfile = NULL, retimg = FALSE, reorient = FALSE,
intern = TRUE, kopts = "", opts = "", verbose = TRUE, ...)
}
\arguments{
\item{file}{(character) image to be dilated}
\item{outfile}{(character) resultant dilated image name}
\item{retimg}{(logical) return image of class nifti}
\item{reorient}{(logical) If retimg, should file be reoriented when read in?
Passed to \code{\link{readNIfTI}}.}
\item{intern}{(logical) to be passed to \code{\link{system}}}
\item{kopts}{(character) options for kernel}
\item{opts}{(character) additional options to be passed to fslmaths}
\item{verbose}{(logical) print out command before running}
\item{...}{additional arguments passed to \code{\link{readNIfTI}}.}
}
\value{
Result from system command, depends if intern is TRUE or FALSE. If
retimg is TRUE, then the image will be returned.
}
\description{
This function calls \code{fslmaths -ero} after inverting the image
to dilate an image with either
the default FSL kernel or the kernel specified in \code{kopts}. The function
either saves the image or returns an object of class nifti.
}
\examples{
if (have.fsl()){
system.time({
x = array(rnorm(1e6), dim = c(100, 100, 100))
img = nifti(x, dim= c(100, 100, 100),
datatype = convert.datatype()$FLOAT32, cal.min = min(x),
cal.max = max(x), pixdim = rep(1, 4))
mask = img > .5
dilated = fsldilate(mask, kopts = "-kernel boxv 5", retimg=TRUE)
})
}
}
| /man/fsldilate.Rd | no_license | emsweene/fslr | R | false | false | 1,583 | rd | % Generated by roxygen2 (4.1.0): do not edit by hand
% Please edit documentation in R/fsldilate.R
\name{fsldilate}
\alias{fsldilate}
\title{Dilate image using FSL}
\usage{
fsldilate(file, outfile = NULL, retimg = FALSE, reorient = FALSE,
intern = TRUE, kopts = "", opts = "", verbose = TRUE, ...)
}
\arguments{
\item{file}{(character) image to be dilated}
\item{outfile}{(character) resultant dilated image name}
\item{retimg}{(logical) return image of class nifti}
\item{reorient}{(logical) If retimg, should file be reoriented when read in?
Passed to \code{\link{readNIfTI}}.}
\item{intern}{(logical) to be passed to \code{\link{system}}}
\item{kopts}{(character) options for kernel}
\item{opts}{(character) additional options to be passed to fslmaths}
\item{verbose}{(logical) print out command before running}
\item{...}{additional arguments passed to \code{\link{readNIfTI}}.}
}
\value{
Result from system command, depends if intern is TRUE or FALSE. If
retimg is TRUE, then the image will be returned.
}
\description{
This function calls \code{fslmaths -ero} after inverting the image
to dilate an image with either
the default FSL kernel or the kernel specified in \code{kopts}. The function
either saves the image or returns an object of class nifti.
}
\examples{
if (have.fsl()){
system.time({
x = array(rnorm(1e6), dim = c(100, 100, 100))
img = nifti(x, dim= c(100, 100, 100),
datatype = convert.datatype()$FLOAT32, cal.min = min(x),
cal.max = max(x), pixdim = rep(1, 4))
mask = img > .5
dilated = fsldilate(mask, kopts = "-kernel boxv 5", retimg=TRUE)
})
}
}
|
# update package if necessary ----
devtools::install_github("kassandra-ru/kassandr")
# load packages ----
library(docxtractr)
library(kassandr)
library(tidyverse)
library(rio)
library(lubridate)
# set data folder ----
info = Sys.info() # ΠΏΠΎΠ»ΡΡΠ°Π΅ΠΌ ΠΈΠ½ΡΠΎΡΠΌΠ°ΡΠΈΡ ΠΎ ΡΠΈΡΡΠ΅ΠΌΠ΅
if (info[1] == "Linux") {
set_libreoffice_path("/usr/bin/libreoffice") # ubuntu or macos
Sys.setenv(LD_LIBRARY_PATH = "/usr/lib/libreoffice/program/") # ubuntu protection against libreglo.so not found
path = "~/Documents/kassandra/data/raw/"
}
if (info[1] == "Windows") {
Sys.setenv("TAR" = "internal") # if install_github() fails on Windows OS
set_libreoffice_path("C:/Program Files/LibreOffice/program/soffice.exe") # windows
path = "D:/Research/Kassandra/data/raw/"
}
# create today's folder ----
access_date = Sys.Date()
today_folder = paste0(path, access_date, "/")
if (!dir.exists(today_folder)) {
dir.create(today_folder)
}
# download setup ----
method = "curl" # maybe "curl", "wget", "libcurl", "auto", "internal", "wininet"
extra = "-L" # options for downloading files, passed to `download.file()`: used for "wget" and "curl" methods
# i_ipc.xlsx ----
url_from = "https://rosstat.gov.ru/storage/mediabank/pddei1ud/i_ipc.xlsx"
raw_path_to = "i_ipc.xlsx"
csv_path_to = "i_ipc.csv"
univariate = TRUE
frequency = 12
comment = "Monthly chained CPI from Russian Statistical Agency"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_i_ipc_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# tab5a.xls ----
url_from = "https://gks.ru/storage/mediabank/e6uKSphi/tab5a.xls"
raw_path_to = "tab5a.xls"
csv_path_to = "tab5a.csv"
univariate = TRUE
frequency = 4
comment = "Gross domestic product quarterly current prices"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_tab5a_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# tab9a.xls ----
url_from = "http://www.gks.ru/free_doc/new_site/vvp/kv/tab9a.xls"
raw_path_to = "tab9a.xls"
csv_path_to = "tab9a.csv"
univariate = TRUE
frequency = 4
comment = "Deflator index in percent to the previous quarter"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_tab9a_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# tab9.xls ----
url_from = "http://www.gks.ru/free_doc/new_site/vvp/kv/tab9.xls"
raw_path_to = "tab9.xls"
csv_path_to = "tab9.csv"
univariate = TRUE
frequency = 4
comment = "Deflator index in percent to the previous quarter early data"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_tab9_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# tab6b.xls ----
url_from = "https://gks.ru/storage/mediabank/35KdpOvs/tab6b.xls"
raw_path_to = "tab6b.xls"
csv_path_to = "tab6b.csv"
univariate = TRUE
frequency = 4
comment = "Gross domestic product quarterly 2016 prices"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_tab6b_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# lendrate.html ----
url_from = "http://www.cbr.ru/hd_base/mkr/mkr_monthes/?UniDbQuery.Posted=True&UniDbQuery.From=08.2000&UniDbQuery.To=01.2100&UniDbQuery.st=SF&UniDbQuery.st=HR&UniDbQuery.st=MB&UniDbQuery.Currency=-1&UniDbQuery.sk=Dd1_&UniDbQuery.sk=Dd7&UniDbQuery.sk=Dd30&UniDbQuery.sk=Dd90&UniDbQuery.sk=Dd180&UniDbQuery.sk=Dd360"
raw_path_to = "lendrate.html"
csv_path_to = "lendrate.csv"
univariate = FALSE
frequency = 12
comment = "Monthly lending rate multiple duration periods"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_lendrate(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# urov_12kv.doc ----
url_from = "http://www.gks.ru/free_doc/new_site/population/urov/urov_12kv.doc"
raw_path_to = "urov_12kv.doc"
csv_path_to = "urov_12kv.csv"
univariate = FALSE
frequency = 12
comment = "Real disposable income percentage to previous period and to same month of previous year"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_urov_12kv_doc(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# 1-07.xlsx ----
url_from = "https://www.gks.ru/bgd/regl/b20_02/IssWWW.exe/Stg/d010/1-07.xlsx"
raw_path_to = "1-07.xlsx"
csv_path_to = "1-07.csv"
univariate = TRUE
frequency = 12
comment = "Construction"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_1_nn_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# 1-03.xlsx ----
url_from = "https://www.gks.ru/bgd/regl/b20_02/IssWWW.exe/Stg/d010/1-03.xlsx"
raw_path_to = "1-03.xlsx"
csv_path_to = "1-03.csv"
univariate = TRUE
frequency = 12
comment = "Agriculture index"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_1_nn_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# 1-11.xlsx ----
url_from = "https://www.gks.ru/bgd/regl/b20_02/IssWWW.exe/Stg/d010/1-11.xlsx"
raw_path_to = "1-11.xlsx"
csv_path_to = "1-11.csv"
univariate = TRUE
frequency = 12
comment = "Budget"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_1_nn_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# m2-m2_sa.xlsx ----
url_from = "http://www.cbr.ru/vfs/statistics/credit_statistics/M2-M2_SA.xlsx"
raw_path_to = "m2-m2_sa.xlsx"
csv_path_to = "m2-m2_sa.csv"
univariate = TRUE
frequency = 12
comment = "Seasonally adjusted M2"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_m2_m2_sa_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# reserves.html ----
url_from = "http://www.cbr.ru/hd_base/mrrf/mrrf_m/?UniDbQuery.Posted=True&UniDbQuery.From=01.1900&UniDbQuery.To=01.2021"
raw_path_to = "reserves.html"
csv_path_to = "reserves.csv"
univariate = FALSE
frequency = 12
comment = "Reserves data from cbr"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_reserves(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# ind_okved2.xlsx ----
url_from = "http://www.gks.ru/free_doc/new_site/business/prom/ind_okved2.xlsx"
raw_path_to = "ind_okved2.xlsx"
csv_path_to = "ind_okved2.csv"
univariate = FALSE
frequency = NA
comment = "Industrial production index"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_ind_okved2_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# trade.xls ----
url_from = "https://www.cbr.ru/vfs/statistics/credit_statistics/trade/trade.xls"
raw_path_to = "trade.xls"
csv_path_to = "trade.csv"
univariate = FALSE
frequency = 12
comment = "Trade statistics"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_trade_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# 1-06-0.xlsx ----
url_from = "http://www.gks.ru/bgd/regl/b20_02/IssWWW.exe/Stg/d010/1-06-0.xlsx"
raw_path_to = "1-06-0.xlsx"
csv_path_to = "invest.csv"
univariate = TRUE
frequency = 4
comment = "Fixed capital investment"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_1_06_0_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# exchangerate.csv ----
csv_path_to = "exchangerate.csv"
univariate = TRUE
frequency = NA
comment = "NA Exchange rate from cbr"
csv_path_to_full = paste0(today_folder, csv_path_to)
data_processed = parse_exchangerate(access_date)
export_with_safe_date(data_processed, csv_path_to_full)
# ind_baza_2018.xls ----
url_from = "https://rosstat.gov.ru/storage/mediabank/BYkjy3Bn/Ind_sub-2018.xls"
raw_path_to = "Ind_sub-2018.xls"
csv_path_to = "ind_baza_2018.csv"
univariate = FALSE
frequency = NA
comment = "Industrial production index, new edition, base year = 2018"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_ind_okved2_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
| /2021-01-25/downloader_v2.R | no_license | kassandra-ru/model | R | false | false | 15,021 | r | # update package if necessary ----
devtools::install_github("kassandra-ru/kassandr")
# load packages ----
library(docxtractr)
library(kassandr)
library(tidyverse)
library(rio)
library(lubridate)
# set data folder ----
info = Sys.info() # ΠΏΠΎΠ»ΡΡΠ°Π΅ΠΌ ΠΈΠ½ΡΠΎΡΠΌΠ°ΡΠΈΡ ΠΎ ΡΠΈΡΡΠ΅ΠΌΠ΅
if (info[1] == "Linux") {
set_libreoffice_path("/usr/bin/libreoffice") # ubuntu or macos
Sys.setenv(LD_LIBRARY_PATH = "/usr/lib/libreoffice/program/") # ubuntu protection against libreglo.so not found
path = "~/Documents/kassandra/data/raw/"
}
if (info[1] == "Windows") {
Sys.setenv("TAR" = "internal") # if install_github() fails on Windows OS
set_libreoffice_path("C:/Program Files/LibreOffice/program/soffice.exe") # windows
path = "D:/Research/Kassandra/data/raw/"
}
# create today's folder ----
access_date = Sys.Date()
today_folder = paste0(path, access_date, "/")
if (!dir.exists(today_folder)) {
dir.create(today_folder)
}
# download setup ----
method = "curl" # maybe "curl", "wget", "libcurl", "auto", "internal", "wininet"
extra = "-L" # options for downloading files, passed to `download.file()`: used for "wget" and "curl" methods
# i_ipc.xlsx ----
url_from = "https://rosstat.gov.ru/storage/mediabank/pddei1ud/i_ipc.xlsx"
raw_path_to = "i_ipc.xlsx"
csv_path_to = "i_ipc.csv"
univariate = TRUE
frequency = 12
comment = "Monthly chained CPI from Russian Statistical Agency"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_i_ipc_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# tab5a.xls ----
url_from = "https://gks.ru/storage/mediabank/e6uKSphi/tab5a.xls"
raw_path_to = "tab5a.xls"
csv_path_to = "tab5a.csv"
univariate = TRUE
frequency = 4
comment = "Gross domestic product quarterly current prices"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_tab5a_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# tab9a.xls ----
url_from = "http://www.gks.ru/free_doc/new_site/vvp/kv/tab9a.xls"
raw_path_to = "tab9a.xls"
csv_path_to = "tab9a.csv"
univariate = TRUE
frequency = 4
comment = "Deflator index in percent to the previous quarter"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_tab9a_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# tab9.xls ----
url_from = "http://www.gks.ru/free_doc/new_site/vvp/kv/tab9.xls"
raw_path_to = "tab9.xls"
csv_path_to = "tab9.csv"
univariate = TRUE
frequency = 4
comment = "Deflator index in percent to the previous quarter early data"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_tab9_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# tab6b.xls ----
url_from = "https://gks.ru/storage/mediabank/35KdpOvs/tab6b.xls"
raw_path_to = "tab6b.xls"
csv_path_to = "tab6b.csv"
univariate = TRUE
frequency = 4
comment = "Gross domestic product quarterly 2016 prices"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_tab6b_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# lendrate.html ----
url_from = "http://www.cbr.ru/hd_base/mkr/mkr_monthes/?UniDbQuery.Posted=True&UniDbQuery.From=08.2000&UniDbQuery.To=01.2100&UniDbQuery.st=SF&UniDbQuery.st=HR&UniDbQuery.st=MB&UniDbQuery.Currency=-1&UniDbQuery.sk=Dd1_&UniDbQuery.sk=Dd7&UniDbQuery.sk=Dd30&UniDbQuery.sk=Dd90&UniDbQuery.sk=Dd180&UniDbQuery.sk=Dd360"
raw_path_to = "lendrate.html"
csv_path_to = "lendrate.csv"
univariate = FALSE
frequency = 12
comment = "Monthly lending rate multiple duration periods"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_lendrate(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# urov_12kv.doc ----
url_from = "http://www.gks.ru/free_doc/new_site/population/urov/urov_12kv.doc"
raw_path_to = "urov_12kv.doc"
csv_path_to = "urov_12kv.csv"
univariate = FALSE
frequency = 12
comment = "Real disposable income percentage to previous period and to same month of previous year"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_urov_12kv_doc(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# 1-07.xlsx ----
url_from = "https://www.gks.ru/bgd/regl/b20_02/IssWWW.exe/Stg/d010/1-07.xlsx"
raw_path_to = "1-07.xlsx"
csv_path_to = "1-07.csv"
univariate = TRUE
frequency = 12
comment = "Construction"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_1_nn_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# 1-03.xlsx ----
url_from = "https://www.gks.ru/bgd/regl/b20_02/IssWWW.exe/Stg/d010/1-03.xlsx"
raw_path_to = "1-03.xlsx"
csv_path_to = "1-03.csv"
univariate = TRUE
frequency = 12
comment = "Agriculture index"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_1_nn_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# 1-11.xlsx ----
url_from = "https://www.gks.ru/bgd/regl/b20_02/IssWWW.exe/Stg/d010/1-11.xlsx"
raw_path_to = "1-11.xlsx"
csv_path_to = "1-11.csv"
univariate = TRUE
frequency = 12
comment = "Budget"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_1_nn_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# m2-m2_sa.xlsx ----
url_from = "http://www.cbr.ru/vfs/statistics/credit_statistics/M2-M2_SA.xlsx"
raw_path_to = "m2-m2_sa.xlsx"
csv_path_to = "m2-m2_sa.csv"
univariate = TRUE
frequency = 12
comment = "Seasonally adjusted M2"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_m2_m2_sa_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# reserves.html ----
url_from = "http://www.cbr.ru/hd_base/mrrf/mrrf_m/?UniDbQuery.Posted=True&UniDbQuery.From=01.1900&UniDbQuery.To=01.2021"
raw_path_to = "reserves.html"
csv_path_to = "reserves.csv"
univariate = FALSE
frequency = 12
comment = "Reserves data from cbr"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_reserves(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# ind_okved2.xlsx ----
url_from = "http://www.gks.ru/free_doc/new_site/business/prom/ind_okved2.xlsx"
raw_path_to = "ind_okved2.xlsx"
csv_path_to = "ind_okved2.csv"
univariate = FALSE
frequency = NA
comment = "Industrial production index"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_ind_okved2_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# trade.xls ----
url_from = "https://www.cbr.ru/vfs/statistics/credit_statistics/trade/trade.xls"
raw_path_to = "trade.xls"
csv_path_to = "trade.csv"
univariate = FALSE
frequency = 12
comment = "Trade statistics"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_trade_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# 1-06-0.xlsx ----
url_from = "http://www.gks.ru/bgd/regl/b20_02/IssWWW.exe/Stg/d010/1-06-0.xlsx"
raw_path_to = "1-06-0.xlsx"
csv_path_to = "invest.csv"
univariate = TRUE
frequency = 4
comment = "Fixed capital investment"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_1_06_0_xlsx(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
# exchangerate.csv ----
csv_path_to = "exchangerate.csv"
univariate = TRUE
frequency = NA
comment = "NA Exchange rate from cbr"
csv_path_to_full = paste0(today_folder, csv_path_to)
data_processed = parse_exchangerate(access_date)
export_with_safe_date(data_processed, csv_path_to_full)
# ind_baza_2018.xls ----
url_from = "https://rosstat.gov.ru/storage/mediabank/BYkjy3Bn/Ind_sub-2018.xls"
raw_path_to = "Ind_sub-2018.xls"
csv_path_to = "ind_baza_2018.csv"
univariate = FALSE
frequency = NA
comment = "Industrial production index, new edition, base year = 2018"
csv_path_to_full = paste0(today_folder, csv_path_to)
raw_path_to_full = paste0(today_folder, raw_path_to)
utils::download.file(url = url_from, destfile = raw_path_to_full, method = method, extra = extra)
if (length(grep("ΠΠΎΡΡΡΠΏ Π·Π°ΠΏΡΠ΅ΡΠ΅Π½", read_lines(raw_path_to_full))) > 0) {
warning("Probably file moved to another location")
stop("Fucking `Access denied` inside a file :(")
}
data_processed = convert_ind_okved2_xls(raw_path_to_full, access_date)
export_with_safe_date(data_processed, csv_path_to_full)
if (file.exists(raw_path_to_full)) {
file.remove(raw_path_to_full)
}
|
## These are a pair of functions that cache the inverse of a matrix. Matrix inversion is usually a costly computation and caching the inverse of a matrix can prove to be efficient for large matrices.
## Usage: If x is the matrix inverse of which to be calculated; cachedmatrix <- makeCacheMatrix(x). Then call cacheSolve on cachedmatrix: cacheSolve(cachedmatrix).
## This function creates a special "matrix" object that can cache its inverse.
makeCacheMatrix <- function(x = matrix()) {
inv <- NULL
set <- function(y){
x <<- y
inv <<- NULL
}
get <- function() x
setinv <- function(solve) inv <<- solve
getinv <- function() inv
list(set = set, get = get, setinv = setinv, getinv = getinv)
}
## This function computes the inverse of the special "matrix" returned by makeCacheMatrix above. If the inverse has already been calculated (and the matrix has not changed), then the cachesolve should retrieve the inverse from the cache.
cacheSolve <- function(x, ...) {
Β Β Β Β Β Β Β Β inv <- x$getinv()
if(!is.null(inv)){
message("getting cached data")
return(inv)
}
data <- x$get()
inv <- solve(data, ...)
x$setinv(inv)
inv
} | /cachematrix.R | no_license | emelaktas/ProgrammingAssignment2 | R | false | false | 1,153 | r | ## These are a pair of functions that cache the inverse of a matrix. Matrix inversion is usually a costly computation and caching the inverse of a matrix can prove to be efficient for large matrices.
## Usage: If x is the matrix inverse of which to be calculated; cachedmatrix <- makeCacheMatrix(x). Then call cacheSolve on cachedmatrix: cacheSolve(cachedmatrix).
## This function creates a special "matrix" object that can cache its inverse.
makeCacheMatrix <- function(x = matrix()) {
inv <- NULL
set <- function(y){
x <<- y
inv <<- NULL
}
get <- function() x
setinv <- function(solve) inv <<- solve
getinv <- function() inv
list(set = set, get = get, setinv = setinv, getinv = getinv)
}
## This function computes the inverse of the special "matrix" returned by makeCacheMatrix above. If the inverse has already been calculated (and the matrix has not changed), then the cachesolve should retrieve the inverse from the cache.
cacheSolve <- function(x, ...) {
Β Β Β Β Β Β Β Β inv <- x$getinv()
if(!is.null(inv)){
message("getting cached data")
return(inv)
}
data <- x$get()
inv <- solve(data, ...)
x$setinv(inv)
inv
} |
# definitions of groups of variables:
var_smd = c("origin", "site", "quadrat", "latitude", "longitude")
var_defense = c("tannin", "lignin", "saponins", "flavonoid", "trichome", "cn", "sla", "ssl")
var_clim = c("MAT", "MAXT", "MINT", "ST", "AP", "WP", "DP", "SP")
var_clim = c("CPC1", "CPC2")
var_symb = c("infection", "shannon")
var_symb = c()
var_herb = c("herbivory")
var_all = c(var_smd, var_defense, var_clim, var_symb, var_herb)
var_def = c(var_smd, var_defense)
load_data = function() {
col_spec = cols(
.default = col_double(),
origin = col_character(),
site = col_character(),
quadrat = col_character()
)
data_native = read_csv("../data/data_native.csv", col_types = col_spec)
data_intro = read_csv("../data/data_intro.csv", col_types = col_spec)
data_full = bind_rows(data_native, data_intro)
data_full = data_full %>% mutate(ssl=.$stemlength/.$stembiomass)
tpc = rda(data_full %>% select(c("MAT","MAXT","MINT","ST")) %>% scale())
ppc = rda(data_full %>% select(c("AP","WP","DP","SP")) %>% scale())
cpc = rda(data_full %>% select(c("MAT", "MAXT", "MINT", "ST", "AP", "WP", "DP", "SP")) %>% scale())
data_full$TPC = tpc$CA$u[,1]
data_full$PPC = ppc$CA$u[,1]
data_full$CPC1 = cpc$CA$u[,1]
data_full$CPC2 = cpc$CA$u[,2]
data_full
}
draw_climate_pca = function(data_full) {
cpc = data_full %>%
select(c("MAT", "MAXT", "MINT", "ST", "AP", "WP", "DP", "SP")) %>%
scale() %>%
rda()
summ.cpc = summary(cpc)
st = summ.cpc$site %>% as.data.frame()
sp = summ.cpc$species %>% as.data.frame()
gr = data.frame(syndrome=data_full$site, origin=data_full$origin)
pca_defense = cpc
fig = ggplot() +
geom_point(aes(st$PC1, st$PC2, shape=gr$origin), size=2) +
geom_segment(aes(x=0, y=0, xend=sp$PC1, yend=sp$PC2),
arrow=arrow(angle=22.5, length=unit(0.2,"cm"), type="closed"),
linetype=1, size=0.5, colour = "gray") +
geom_text(aes(sp$PC1, sp$PC2, label=row.names(sp)), hjust=0.2, vjust=1.5) +
labs(x=paste("PC1 (", format(100 * summary(pca_defense)$cont[[1]][2,1], digits=4), "%)", sep=""),
y=paste("PC2 (", format(100 * summary(pca_defense)$cont[[1]][2,2], digits=4), "%)", sep="")) +
geom_hline(yintercept=0,linetype=2,size=0.2) +
geom_vline(xintercept=0,linetype=2,size=0.2)+
guides(shape=guide_legend(title="Origin"),color=guide_legend(title="Origin")) +
scale_shape_manual(values = c(16, 1)) +
theme_bw()
fig
}
draw_defense_pca = function(st, sp, gr, pca_defense) {
ggplot() +
geom_point(aes(st$PC1, st$PC2, color=gr$syndrome, shape=gr$origin)) +
stat_ellipse(aes(st$PC1, st$PC2, color=gr$syndrome, group=gr$syndrome)) +
geom_segment(aes(x=0, y=0, xend=sp$PC1, yend=sp$PC2),
arrow=arrow(angle=22.5, length=unit(0.2,"cm"), type="closed"),
linetype=1, size=0.5, colour = "red") +
geom_text(aes(sp$PC1, sp$PC2, label=row.names(sp)), hjust=0.2, vjust=1.5) +
labs(x=paste("PC1 (", format(100 * summary(pca_defense)$cont[[1]][2,1], digits=4), "%)", sep=""),
y=paste("PC2 (", format(100 * summary(pca_defense)$cont[[1]][2,2], digits=4), "%)", sep="")) +
geom_hline(yintercept=0,linetype=2,size=0.2) +
geom_vline(xintercept=0,linetype=2,size=0.2)+
guides(shape=guide_legend(title="Origin"),color=guide_legend(title="Syndrome")) +
theme_bw()
}
clust_and_pca = function(data) {
hcl = data %>%
select(var_defense) %>%
scale() %>%
dist() %>%
hclust("ward.D")
gr = as.factor(cutree(hcl, 4))
pca_defense = data %>%
select(var_defense) %>%
scale() %>%
rda()
st = summary(pca_defense)$sites %>% as.data.frame()
sp = summary(pca_defense)$species %>% as.data.frame()
gr = data.frame(syndrome=gr, origin=data$origin)
list(st=st, sp=sp, gr=gr, pca_defense=pca_defense)
}
permutest_cluster = function(data_def_nona, lst) {
x = data_def_nona %>%
mutate(syndrome=lst$gr$syndrome)
x_s12 = filter(x, syndrome %in% c(1, 2))
x_s13 = filter(x, syndrome %in% c(1, 3))
x_s14 = filter(x, syndrome %in% c(1, 4))
x_s23 = filter(x, syndrome %in% c(2, 3))
x_s24 = filter(x, syndrome %in% c(2, 4))
x_s34 = filter(x, syndrome %in% c(3, 4))
# multi-groups
multi = adonis(select(x, var_defense) ~ syndrome, data=x)
# parse-wise
pw1 = adonis(select(x_s12, var_defense) ~ syndrome, data=x_s12)
pw2 = adonis(select(x_s13, var_defense) ~ syndrome, data=x_s13)
pw3 = adonis(select(x_s14, var_defense) ~ syndrome, data=x_s14)
pw4 = adonis(select(x_s23, var_defense) ~ syndrome, data=x_s23)
pw5 = adonis(select(x_s24, var_defense) ~ syndrome, data=x_s24)
pw6 = adonis(select(x_s34, var_defense) ~ syndrome, data=x_s34)
rbind(multi$aov.tab, pw1$aov.tab, pw2$aov.tab, pw3$aov.tab, pw4$aov.tab, pw5$aov.tab, pw6$aov.tab)
}
centroids_and_area = function(data_def_nona, data_all_nona) {
x = data_def_nona %>%
mutate(syndrome=lst$gr$syndrome) %>%
select(var_defense) %>%
scale()
center = sweep(x, 2, apply(x, 2, min),'-')
R = apply(x, 2, max) - apply(x, 2, min)
x_star = sweep(center, 2, R, "/")
x_star = as_tibble(x_star)[,c(3,6,2,7,8,5,1,4)]
theta = (c(1,2,3,4,5,6,7,8)-1) * 2*pi/8
xm = sweep(x_star, 2, cos(theta), "*")
ym = sweep(x_star, 2, sin(theta), "*")
centroids = t(apply(cbind(xm, ym), 1, function(x) pracma::poly_center(x[1:8], x[9:16])))
area = apply(cbind(xm, ym), 1, function(x) pracma::polyarea(x[1:8], x[9:16]))
df = tibble(
area = area,
cent.x = centroids[,1],
cent.y = centroids[,2],
syndrome = lst$gr$syndrome,
quadrat = data_def_nona$quadrat
)
data = left_join(data_all_nona, df, by="quadrat")
data
}
# Mantel test
multiple_mantel = function(data, prefix) {
dmat_defense = data %>% select(var_defense) %>% scale() %>% dist()
dmat_cent = data %>% select(cent.x, cent.y) %>% scale() %>% dist()
dmat_area = data %>% select(area) %>% scale() %>% dist()
dmat_clim = data %>% select(var_clim) %>% scale() %>% dist()
dmat_symb = data %>% select(var_symb) %>% scale()%>% dist()
dmat_herb = data %>% select(var_herb) %>% scale() %>% dist()
mm_d = multi.mantel(dmat_defense, list(clim=dmat_clim, symb=dmat_symb, herb=dmat_herb))
mm_c = multi.mantel(dmat_cent, list(clim=dmat_clim, symb=dmat_symb, herb=dmat_herb))
mm_a = multi.mantel(dmat_area, list(clim=dmat_clim, symb=dmat_symb, herb=dmat_herb))
df1 = cbind(coeff=mm_d$coefficients, t=mm_d$tstatistic, p=mm_d$probt, rsq=mm_d$r.squared)[2:4,]
rownames(df1) = c("CLIM", "SYMB", "HERB")
df2 = cbind(coeff=mm_a$coefficients, t=mm_a$tstatistic, p=mm_a$probt, rsq=mm_a$r.squared)[2:4,]
rownames(df2) = c("CLIM", "SYMB", "HERB")
df3 = cbind(coeff=mm_c$coefficients, t=mm_c$tstatistic, p=mm_c$probt, rsq=mm_c$r.squared)[2:4,]
rownames(df3) = c("CLIM", "SYMB", "HERB")
write.csv(rbind(df1,df2,df3), paste0("../",prefix,"_mantel.csv"))
}
rda_and_summ = function(data, var, fname) {
rvar = data %>% select(var) %>% scale() %>% as.data.frame()
evar = cbind(
data %>% select(var_clim) %>% scale(),
data %>% select(var_herb) %>% scale(),
data %>% select(origin, latitude)
) %>% as.data.frame()
ret = rda(
rvar ~
origin*latitude,
data = evar
)
perm <- how(nperm = 999)
#setBlocks(perm) <- with(data, site)
aov.rda = anova(ret, by="term", permutations = perm)
write.csv(aov.rda, fname)
aov.rda
}
vp_and_plot_EOL = function(data, var, fname) {
rvar = data %>% select(var) %>% scale() %>% as.data.frame()
env = cbind(
data %>% select(var_clim) %>% scale(),
data %>% select(var_symb) %>% scale(),
data %>% select(var_herb) %>% scale()
)
org = data %>% select(origin)
lat = data %>% select(latitude)
vp = varpart(rvar, env, lat, org)
plot(vp, Xnames=c("ENV","LAT","ORG"))
title(fname)
}
vp_and_plot_ABO = function(data, var, fname) {
rvar = data %>% select(var) %>% scale() %>% as.data.frame()
bio = cbind(
data %>% select(var_symb) %>% scale(),
data %>% select(var_herb) %>% scale()
)
abio = data %>% select(var_clim) %>% scale()
org = data %>% select(origin)
#lat = data %>% select(latitude)
vp = varpart(rvar, bio, abio, org)
plot(vp, Xnames=c("BIO","ABIO","ORG"))
title(fname)
} | /code/utils.R | permissive | Augustpan/Latitudinal_Defense_Syndromes | R | false | false | 8,268 | r | # definitions of groups of variables:
var_smd = c("origin", "site", "quadrat", "latitude", "longitude")
var_defense = c("tannin", "lignin", "saponins", "flavonoid", "trichome", "cn", "sla", "ssl")
var_clim = c("MAT", "MAXT", "MINT", "ST", "AP", "WP", "DP", "SP")
var_clim = c("CPC1", "CPC2")
var_symb = c("infection", "shannon")
var_symb = c()
var_herb = c("herbivory")
var_all = c(var_smd, var_defense, var_clim, var_symb, var_herb)
var_def = c(var_smd, var_defense)
load_data = function() {
col_spec = cols(
.default = col_double(),
origin = col_character(),
site = col_character(),
quadrat = col_character()
)
data_native = read_csv("../data/data_native.csv", col_types = col_spec)
data_intro = read_csv("../data/data_intro.csv", col_types = col_spec)
data_full = bind_rows(data_native, data_intro)
data_full = data_full %>% mutate(ssl=.$stemlength/.$stembiomass)
tpc = rda(data_full %>% select(c("MAT","MAXT","MINT","ST")) %>% scale())
ppc = rda(data_full %>% select(c("AP","WP","DP","SP")) %>% scale())
cpc = rda(data_full %>% select(c("MAT", "MAXT", "MINT", "ST", "AP", "WP", "DP", "SP")) %>% scale())
data_full$TPC = tpc$CA$u[,1]
data_full$PPC = ppc$CA$u[,1]
data_full$CPC1 = cpc$CA$u[,1]
data_full$CPC2 = cpc$CA$u[,2]
data_full
}
draw_climate_pca = function(data_full) {
cpc = data_full %>%
select(c("MAT", "MAXT", "MINT", "ST", "AP", "WP", "DP", "SP")) %>%
scale() %>%
rda()
summ.cpc = summary(cpc)
st = summ.cpc$site %>% as.data.frame()
sp = summ.cpc$species %>% as.data.frame()
gr = data.frame(syndrome=data_full$site, origin=data_full$origin)
pca_defense = cpc
fig = ggplot() +
geom_point(aes(st$PC1, st$PC2, shape=gr$origin), size=2) +
geom_segment(aes(x=0, y=0, xend=sp$PC1, yend=sp$PC2),
arrow=arrow(angle=22.5, length=unit(0.2,"cm"), type="closed"),
linetype=1, size=0.5, colour = "gray") +
geom_text(aes(sp$PC1, sp$PC2, label=row.names(sp)), hjust=0.2, vjust=1.5) +
labs(x=paste("PC1 (", format(100 * summary(pca_defense)$cont[[1]][2,1], digits=4), "%)", sep=""),
y=paste("PC2 (", format(100 * summary(pca_defense)$cont[[1]][2,2], digits=4), "%)", sep="")) +
geom_hline(yintercept=0,linetype=2,size=0.2) +
geom_vline(xintercept=0,linetype=2,size=0.2)+
guides(shape=guide_legend(title="Origin"),color=guide_legend(title="Origin")) +
scale_shape_manual(values = c(16, 1)) +
theme_bw()
fig
}
draw_defense_pca = function(st, sp, gr, pca_defense) {
ggplot() +
geom_point(aes(st$PC1, st$PC2, color=gr$syndrome, shape=gr$origin)) +
stat_ellipse(aes(st$PC1, st$PC2, color=gr$syndrome, group=gr$syndrome)) +
geom_segment(aes(x=0, y=0, xend=sp$PC1, yend=sp$PC2),
arrow=arrow(angle=22.5, length=unit(0.2,"cm"), type="closed"),
linetype=1, size=0.5, colour = "red") +
geom_text(aes(sp$PC1, sp$PC2, label=row.names(sp)), hjust=0.2, vjust=1.5) +
labs(x=paste("PC1 (", format(100 * summary(pca_defense)$cont[[1]][2,1], digits=4), "%)", sep=""),
y=paste("PC2 (", format(100 * summary(pca_defense)$cont[[1]][2,2], digits=4), "%)", sep="")) +
geom_hline(yintercept=0,linetype=2,size=0.2) +
geom_vline(xintercept=0,linetype=2,size=0.2)+
guides(shape=guide_legend(title="Origin"),color=guide_legend(title="Syndrome")) +
theme_bw()
}
clust_and_pca = function(data) {
hcl = data %>%
select(var_defense) %>%
scale() %>%
dist() %>%
hclust("ward.D")
gr = as.factor(cutree(hcl, 4))
pca_defense = data %>%
select(var_defense) %>%
scale() %>%
rda()
st = summary(pca_defense)$sites %>% as.data.frame()
sp = summary(pca_defense)$species %>% as.data.frame()
gr = data.frame(syndrome=gr, origin=data$origin)
list(st=st, sp=sp, gr=gr, pca_defense=pca_defense)
}
permutest_cluster = function(data_def_nona, lst) {
x = data_def_nona %>%
mutate(syndrome=lst$gr$syndrome)
x_s12 = filter(x, syndrome %in% c(1, 2))
x_s13 = filter(x, syndrome %in% c(1, 3))
x_s14 = filter(x, syndrome %in% c(1, 4))
x_s23 = filter(x, syndrome %in% c(2, 3))
x_s24 = filter(x, syndrome %in% c(2, 4))
x_s34 = filter(x, syndrome %in% c(3, 4))
# multi-groups
multi = adonis(select(x, var_defense) ~ syndrome, data=x)
# parse-wise
pw1 = adonis(select(x_s12, var_defense) ~ syndrome, data=x_s12)
pw2 = adonis(select(x_s13, var_defense) ~ syndrome, data=x_s13)
pw3 = adonis(select(x_s14, var_defense) ~ syndrome, data=x_s14)
pw4 = adonis(select(x_s23, var_defense) ~ syndrome, data=x_s23)
pw5 = adonis(select(x_s24, var_defense) ~ syndrome, data=x_s24)
pw6 = adonis(select(x_s34, var_defense) ~ syndrome, data=x_s34)
rbind(multi$aov.tab, pw1$aov.tab, pw2$aov.tab, pw3$aov.tab, pw4$aov.tab, pw5$aov.tab, pw6$aov.tab)
}
centroids_and_area = function(data_def_nona, data_all_nona) {
x = data_def_nona %>%
mutate(syndrome=lst$gr$syndrome) %>%
select(var_defense) %>%
scale()
center = sweep(x, 2, apply(x, 2, min),'-')
R = apply(x, 2, max) - apply(x, 2, min)
x_star = sweep(center, 2, R, "/")
x_star = as_tibble(x_star)[,c(3,6,2,7,8,5,1,4)]
theta = (c(1,2,3,4,5,6,7,8)-1) * 2*pi/8
xm = sweep(x_star, 2, cos(theta), "*")
ym = sweep(x_star, 2, sin(theta), "*")
centroids = t(apply(cbind(xm, ym), 1, function(x) pracma::poly_center(x[1:8], x[9:16])))
area = apply(cbind(xm, ym), 1, function(x) pracma::polyarea(x[1:8], x[9:16]))
df = tibble(
area = area,
cent.x = centroids[,1],
cent.y = centroids[,2],
syndrome = lst$gr$syndrome,
quadrat = data_def_nona$quadrat
)
data = left_join(data_all_nona, df, by="quadrat")
data
}
# Mantel test
multiple_mantel = function(data, prefix) {
dmat_defense = data %>% select(var_defense) %>% scale() %>% dist()
dmat_cent = data %>% select(cent.x, cent.y) %>% scale() %>% dist()
dmat_area = data %>% select(area) %>% scale() %>% dist()
dmat_clim = data %>% select(var_clim) %>% scale() %>% dist()
dmat_symb = data %>% select(var_symb) %>% scale()%>% dist()
dmat_herb = data %>% select(var_herb) %>% scale() %>% dist()
mm_d = multi.mantel(dmat_defense, list(clim=dmat_clim, symb=dmat_symb, herb=dmat_herb))
mm_c = multi.mantel(dmat_cent, list(clim=dmat_clim, symb=dmat_symb, herb=dmat_herb))
mm_a = multi.mantel(dmat_area, list(clim=dmat_clim, symb=dmat_symb, herb=dmat_herb))
df1 = cbind(coeff=mm_d$coefficients, t=mm_d$tstatistic, p=mm_d$probt, rsq=mm_d$r.squared)[2:4,]
rownames(df1) = c("CLIM", "SYMB", "HERB")
df2 = cbind(coeff=mm_a$coefficients, t=mm_a$tstatistic, p=mm_a$probt, rsq=mm_a$r.squared)[2:4,]
rownames(df2) = c("CLIM", "SYMB", "HERB")
df3 = cbind(coeff=mm_c$coefficients, t=mm_c$tstatistic, p=mm_c$probt, rsq=mm_c$r.squared)[2:4,]
rownames(df3) = c("CLIM", "SYMB", "HERB")
write.csv(rbind(df1,df2,df3), paste0("../",prefix,"_mantel.csv"))
}
rda_and_summ = function(data, var, fname) {
rvar = data %>% select(var) %>% scale() %>% as.data.frame()
evar = cbind(
data %>% select(var_clim) %>% scale(),
data %>% select(var_herb) %>% scale(),
data %>% select(origin, latitude)
) %>% as.data.frame()
ret = rda(
rvar ~
origin*latitude,
data = evar
)
perm <- how(nperm = 999)
#setBlocks(perm) <- with(data, site)
aov.rda = anova(ret, by="term", permutations = perm)
write.csv(aov.rda, fname)
aov.rda
}
vp_and_plot_EOL = function(data, var, fname) {
rvar = data %>% select(var) %>% scale() %>% as.data.frame()
env = cbind(
data %>% select(var_clim) %>% scale(),
data %>% select(var_symb) %>% scale(),
data %>% select(var_herb) %>% scale()
)
org = data %>% select(origin)
lat = data %>% select(latitude)
vp = varpart(rvar, env, lat, org)
plot(vp, Xnames=c("ENV","LAT","ORG"))
title(fname)
}
vp_and_plot_ABO = function(data, var, fname) {
rvar = data %>% select(var) %>% scale() %>% as.data.frame()
bio = cbind(
data %>% select(var_symb) %>% scale(),
data %>% select(var_herb) %>% scale()
)
abio = data %>% select(var_clim) %>% scale()
org = data %>% select(origin)
#lat = data %>% select(latitude)
vp = varpart(rvar, bio, abio, org)
plot(vp, Xnames=c("BIO","ABIO","ORG"))
title(fname)
} |
\name{Huber}
\alias{Huber}
\title{Huber Loss
}
\description{Evaluates the Huber loss function defined as \deqn{ f(r) = \left\{
\begin{array}{ll}
\frac{1}{2}|r|^2 & |r| \le c \\
c(|r|-\frac{1}{2}c) & |r| > c
\end{array}
\right. }{f(r)=(1/2)*r^2 if |r|<=c}
\deqn{}{f(r)=c*(|r|-(1/2)*c) if |r|>c }
}
\usage{
Huber(r, c = 1.345)
}
\arguments{
\item{r }{a real number or vector.
}
\item{c }{a positive number. If the value is negative, it's absolute value will be used.
}
}
\examples{
set.seed(1)
x = rnorm(200, mean = 1)
y = Huber(x)
plot(x, y)
abline(h = (1.345)^2/2)
}
| /man/Huber.Rd | no_license | cran/qrmix | R | false | false | 585 | rd | \name{Huber}
\alias{Huber}
\title{Huber Loss
}
\description{Evaluates the Huber loss function defined as \deqn{ f(r) = \left\{
\begin{array}{ll}
\frac{1}{2}|r|^2 & |r| \le c \\
c(|r|-\frac{1}{2}c) & |r| > c
\end{array}
\right. }{f(r)=(1/2)*r^2 if |r|<=c}
\deqn{}{f(r)=c*(|r|-(1/2)*c) if |r|>c }
}
\usage{
Huber(r, c = 1.345)
}
\arguments{
\item{r }{a real number or vector.
}
\item{c }{a positive number. If the value is negative, it's absolute value will be used.
}
}
\examples{
set.seed(1)
x = rnorm(200, mean = 1)
y = Huber(x)
plot(x, y)
abline(h = (1.345)^2/2)
}
|
\name{xmb}
\alias{xmb}
\title{More robust test of whether something is NA or null.
Should not give an error or NA, whatever x is.}
\usage{
xmb(x, y = "")
}
\arguments{
\item{x}{The object to be tested. May not even exist.}
\item{y}{true if missing or null or y, otherwise false.
NOTE IT GIVES F IF IT IS ANY DATA FRAME, EVEN AN EMPTY
ONE}
}
\description{
More robust test of whether something is NA or null. Should
not give an error or NA, whatever x is.
}
| /man/xmb.Rd | no_license | stevepowell99/omnivr | R | false | false | 467 | rd | \name{xmb}
\alias{xmb}
\title{More robust test of whether something is NA or null.
Should not give an error or NA, whatever x is.}
\usage{
xmb(x, y = "")
}
\arguments{
\item{x}{The object to be tested. May not even exist.}
\item{y}{true if missing or null or y, otherwise false.
NOTE IT GIVES F IF IT IS ANY DATA FRAME, EVEN AN EMPTY
ONE}
}
\description{
More robust test of whether something is NA or null. Should
not give an error or NA, whatever x is.
}
|
dataset <- read.csv("C:/Users/jensb/Documents/dataanalyse/dataanalyse/datasets/dataset-82167.csv")
View(dataset.82167)
# geef de namen van de variabelen weer
names(dataset)
# geef enkel de variabele hd weer
dataset$hd
| /eerstetest.R | no_license | embobrecht/dataanalyse | R | false | false | 225 | r |
dataset <- read.csv("C:/Users/jensb/Documents/dataanalyse/dataanalyse/datasets/dataset-82167.csv")
View(dataset.82167)
# geef de namen van de variabelen weer
names(dataset)
# geef enkel de variabele hd weer
dataset$hd
|
with(a5db1d12f78c845b4a783167e77071f2f, {ROOT <- 'D:/SEMOSS_v4.0.0_x64/SEMOSS_v4.0.0_x64/semosshome/db/Atadata2__3b3e4a3b-d382-4e98-9950-9b4e8b308c1c/version/80bb2a25-ac5d-47d0-abfc-b3f3811f0936';rm(list=ls())}); | /80bb2a25-ac5d-47d0-abfc-b3f3811f0936/R/Temp/aMgSCnlb73PO6.R | no_license | ayanmanna8/test | R | false | false | 212 | r | with(a5db1d12f78c845b4a783167e77071f2f, {ROOT <- 'D:/SEMOSS_v4.0.0_x64/SEMOSS_v4.0.0_x64/semosshome/db/Atadata2__3b3e4a3b-d382-4e98-9950-9b4e8b308c1c/version/80bb2a25-ac5d-47d0-abfc-b3f3811f0936';rm(list=ls())}); |
.libPaths("/data/hydro/R_libs35")
.libPaths()
library(data.table)
library(dplyr)
library(tidyverse)
library(lubridate)
source_path = "/home/hnoorazar/chilling_codes/current_draft/chill_core.R"
source(source_path)
options(digit=9)
options(digits=9)
start_time <- Sys.time()
######################################################################
args = commandArgs(trailingOnly=TRUE)
model_type = args[1]
observed_dt <- data.table()
######################################################################
# Define main output path
frost_out = "/data/hydro/users/Hossein/chill/frost_bloom_initial_database/"
param_dir = file.path("/home/hnoorazar/chilling_codes/parameters/")
local_files <- read.delim(file = paste0(param_dir, "file_list.txt"), header = F)
local_files <- as.vector(local_files$V1)
LocationGroups_NoMontana <- read.csv(paste0(param_dir, "LocationGroups_NoMontana.csv"),
header=T, sep=",", as.is=T)
LocationGroups_NoMontana <- within(LocationGroups_NoMontana, remove(lat, long))
######################################################################
observed_dir <- "/data/hydro/jennylabcommon2/metdata/historical/UI_historical/VIC_Binary_CONUS_to_2016/"
setwd(observed_dir)
print (paste0("we should be in :", observed_dir))
print (getwd())
dir_con <- dir()
dir_con <- dir_con[grep(pattern = "data_", x = dir_con)]
dir_con <- dir_con[which(dir_con %in% local_files)]
# 3. Process the data -----------------------------------------------------
for(file in dir_con){
lat <- substr(x = file, start = 6, stop = 13)
long <- substr(x = file, start = 15, stop = 24)
met_data <- read_binary(file_path = file, hist = T, no_vars=8)
met_data <- data.table(met_data)
# Clean it up
met_data <- met_data %>%
select(c(year, month, day, tmin)) %>%
data.table()
met_data <- form_chill_season_day_of_year_observed(met_data)
met_data$lat <- lat
met_data$long <- long
met_data$model <- "Observed"
observed_dt <- rbind(observed_dt, met_data)
}
observed_dt <- remove_montana(observed_dt, LocationGroups_NoMontana)
observed_dt <- within(observed_dt, remove(lat, long))
observed_dt <- add_time_periods_observed(observed_dt)
observed_dt <- observed_dt %>% filter(tmin <= 0)
observed_dt_till_Dec <- observed_dt %>% filter(month %in% c(9, 10, 11, 12))
observed_dt_till_Jan <- observed_dt %>% filter(month %in% c(9, 10, 11, 12, 13))
observed_dt_till_Feb <- observed_dt
rm(observed_dt)
################################################################
######## Jan of 1950 is in the data, which belongs to chill season 1949,
######## which we do not have it in the data, so, time period for them will be
######## NA, so we drop them.
observed_dt_till_Jan <- na.omit(observed_dt_till_Jan)
observed_dt_till_Feb <- na.omit(observed_dt_till_Feb)
if (dir.exists(frost_out) == F) { dir.create(path = frost_out, recursive = T) }
saveRDS(observed_dt_till_Dec, paste0(frost_out, "observed_dt_till_Dec.rds"))
saveRDS(observed_dt_till_Jan, paste0(frost_out, "observed_dt_till_Jan.rds"))
saveRDS(observed_dt_till_Feb, paste0(frost_out, "observed_dt_till_Feb.rds"))
end_time <- Sys.time()
print( end_time - start_time)
| /chilling/03_frost_bloom_aeolus/d_read_binary_for_frost_obs.R | permissive | HNoorazar/Ag | R | false | false | 3,213 | r | .libPaths("/data/hydro/R_libs35")
.libPaths()
library(data.table)
library(dplyr)
library(tidyverse)
library(lubridate)
source_path = "/home/hnoorazar/chilling_codes/current_draft/chill_core.R"
source(source_path)
options(digit=9)
options(digits=9)
start_time <- Sys.time()
######################################################################
args = commandArgs(trailingOnly=TRUE)
model_type = args[1]
observed_dt <- data.table()
######################################################################
# Define main output path
frost_out = "/data/hydro/users/Hossein/chill/frost_bloom_initial_database/"
param_dir = file.path("/home/hnoorazar/chilling_codes/parameters/")
local_files <- read.delim(file = paste0(param_dir, "file_list.txt"), header = F)
local_files <- as.vector(local_files$V1)
LocationGroups_NoMontana <- read.csv(paste0(param_dir, "LocationGroups_NoMontana.csv"),
header=T, sep=",", as.is=T)
LocationGroups_NoMontana <- within(LocationGroups_NoMontana, remove(lat, long))
######################################################################
observed_dir <- "/data/hydro/jennylabcommon2/metdata/historical/UI_historical/VIC_Binary_CONUS_to_2016/"
setwd(observed_dir)
print (paste0("we should be in :", observed_dir))
print (getwd())
dir_con <- dir()
dir_con <- dir_con[grep(pattern = "data_", x = dir_con)]
dir_con <- dir_con[which(dir_con %in% local_files)]
# 3. Process the data -----------------------------------------------------
for(file in dir_con){
lat <- substr(x = file, start = 6, stop = 13)
long <- substr(x = file, start = 15, stop = 24)
met_data <- read_binary(file_path = file, hist = T, no_vars=8)
met_data <- data.table(met_data)
# Clean it up
met_data <- met_data %>%
select(c(year, month, day, tmin)) %>%
data.table()
met_data <- form_chill_season_day_of_year_observed(met_data)
met_data$lat <- lat
met_data$long <- long
met_data$model <- "Observed"
observed_dt <- rbind(observed_dt, met_data)
}
observed_dt <- remove_montana(observed_dt, LocationGroups_NoMontana)
observed_dt <- within(observed_dt, remove(lat, long))
observed_dt <- add_time_periods_observed(observed_dt)
observed_dt <- observed_dt %>% filter(tmin <= 0)
observed_dt_till_Dec <- observed_dt %>% filter(month %in% c(9, 10, 11, 12))
observed_dt_till_Jan <- observed_dt %>% filter(month %in% c(9, 10, 11, 12, 13))
observed_dt_till_Feb <- observed_dt
rm(observed_dt)
################################################################
######## Jan of 1950 is in the data, which belongs to chill season 1949,
######## which we do not have it in the data, so, time period for them will be
######## NA, so we drop them.
observed_dt_till_Jan <- na.omit(observed_dt_till_Jan)
observed_dt_till_Feb <- na.omit(observed_dt_till_Feb)
if (dir.exists(frost_out) == F) { dir.create(path = frost_out, recursive = T) }
saveRDS(observed_dt_till_Dec, paste0(frost_out, "observed_dt_till_Dec.rds"))
saveRDS(observed_dt_till_Jan, paste0(frost_out, "observed_dt_till_Jan.rds"))
saveRDS(observed_dt_till_Feb, paste0(frost_out, "observed_dt_till_Feb.rds"))
end_time <- Sys.time()
print( end_time - start_time)
|
context("messages")
test_that("messages in callr::r do not crash session", {
ret <- r(function() { cliapp::cli_text("fooobar"); 1 + 1 })
expect_identical(ret, 2)
gc()
})
test_that("messages in callr::r_bg do not crash session", {
skip_in_covr() # TODO: what wrong with this on Windows?
rx <- r_bg(function() { cliapp::cli_text("fooobar"); 1 + 1 })
rx$wait(5000)
rx$kill()
expect_equal(rx$get_exit_status(), 0)
expect_equal(rx$get_result(), 2)
processx::processx_conn_close(rx$get_output_connection())
processx::processx_conn_close(rx$get_error_connection())
gc()
})
| /tests/testthat/test-messages.R | permissive | bedantaguru/callr | R | false | false | 596 | r |
context("messages")
test_that("messages in callr::r do not crash session", {
ret <- r(function() { cliapp::cli_text("fooobar"); 1 + 1 })
expect_identical(ret, 2)
gc()
})
test_that("messages in callr::r_bg do not crash session", {
skip_in_covr() # TODO: what wrong with this on Windows?
rx <- r_bg(function() { cliapp::cli_text("fooobar"); 1 + 1 })
rx$wait(5000)
rx$kill()
expect_equal(rx$get_exit_status(), 0)
expect_equal(rx$get_result(), 2)
processx::processx_conn_close(rx$get_output_connection())
processx::processx_conn_close(rx$get_error_connection())
gc()
})
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/deviationFunctions.R
\name{sortedIndex}
\alias{sortedIndex}
\title{Index map for data set}
\usage{
sortedIndex(dt)
}
\arguments{
\item{dt}{data.table}
}
\value{
A data.table is returned that consists of indexes for the order of each column. E.g. a data table with the first column (0.3,0.5, 0.1) would get an index of (2,3,1), i.e. the elements sorted by (2,3,1) are in increasing order. The index starts with 1.
}
\description{
The index map allows to create subspace slices without working on the underlying data set but rather on the ordered index.
}
| /man/sortedIndex.Rd | no_license | holtri/R-subcon | R | false | true | 632 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/deviationFunctions.R
\name{sortedIndex}
\alias{sortedIndex}
\title{Index map for data set}
\usage{
sortedIndex(dt)
}
\arguments{
\item{dt}{data.table}
}
\value{
A data.table is returned that consists of indexes for the order of each column. E.g. a data table with the first column (0.3,0.5, 0.1) would get an index of (2,3,1), i.e. the elements sorted by (2,3,1) are in increasing order. The index starts with 1.
}
\description{
The index map allows to create subspace slices without working on the underlying data set but rather on the ordered index.
}
|
# Have you ever been on the data.gov.uk and been completely perplexed to where the data is available to download
# ** hand shoots up **
# Well I've just found a brilliant stackoverflow post (https://stackoverflow.com/questions/32034495/list-aviable-wfs-layers-and-read-into-data-frame-with-rgdal)
# that will enable you to get the data that you need into R (providing it is in WFS format).
# Firstly you'll want to go to the webpage where the data resides: https://data.gov.uk/dataset/fife-secondary-school-catchment-areas
# Then you want to click on the link to the WFS file (more info on this format here: https://en.wikipedia.org/wiki/Web_Feature_Service)
# Then copy the link, for example:
# http://arcgisweb.fife.gov.uk/geoserver/fife/ows?service=WFS&version=1.0.0&
# request=GetFeature&typeName=fife:SCHOOL_SECONDARY_CATCHMENTS&maxFeatures=1000000&outputFormat=GML2
# Now all you need to do is run the code provided by 'hrbrmstr', replacing the dsn with the copied link:
library(gdalUtils)
library(rgdal)
dsn <- "WFS:http://arcgisweb.fife.gov.uk/geoserver/fife/ows?service=WFS&version=1.0.0&
request=GetFeature&typeName=fife:SCHOOL_SECONDARY_CATCHMENTS&maxFeatures=1000000&outputFormat=GML2"
ogrinfo(dsn, so=TRUE)
# You need to put in what you want to extract here..
ogr2ogr(dsn, "sic.shp", "fife:SCHOOL_SECONDARY_CATCHMENTS")
sic <- readOGR("sic.shp", "sic", stringsAsFactors=FALSE)
plot(sic)
# Apparently there are 1,729 WFS files on data.gov.uk so hopefully this comes in useful
| /readingWFSfiles.R | no_license | markocherrie/Helpful_Code | R | false | false | 1,505 | r | # Have you ever been on the data.gov.uk and been completely perplexed to where the data is available to download
# ** hand shoots up **
# Well I've just found a brilliant stackoverflow post (https://stackoverflow.com/questions/32034495/list-aviable-wfs-layers-and-read-into-data-frame-with-rgdal)
# that will enable you to get the data that you need into R (providing it is in WFS format).
# Firstly you'll want to go to the webpage where the data resides: https://data.gov.uk/dataset/fife-secondary-school-catchment-areas
# Then you want to click on the link to the WFS file (more info on this format here: https://en.wikipedia.org/wiki/Web_Feature_Service)
# Then copy the link, for example:
# http://arcgisweb.fife.gov.uk/geoserver/fife/ows?service=WFS&version=1.0.0&
# request=GetFeature&typeName=fife:SCHOOL_SECONDARY_CATCHMENTS&maxFeatures=1000000&outputFormat=GML2
# Now all you need to do is run the code provided by 'hrbrmstr', replacing the dsn with the copied link:
library(gdalUtils)
library(rgdal)
dsn <- "WFS:http://arcgisweb.fife.gov.uk/geoserver/fife/ows?service=WFS&version=1.0.0&
request=GetFeature&typeName=fife:SCHOOL_SECONDARY_CATCHMENTS&maxFeatures=1000000&outputFormat=GML2"
ogrinfo(dsn, so=TRUE)
# You need to put in what you want to extract here..
ogr2ogr(dsn, "sic.shp", "fife:SCHOOL_SECONDARY_CATCHMENTS")
sic <- readOGR("sic.shp", "sic", stringsAsFactors=FALSE)
plot(sic)
# Apparently there are 1,729 WFS files on data.gov.uk so hopefully this comes in useful
|
### Initialization ####
library('RPostgres')
library('RPostgreSQL')
### Creating connector
db <- 'INSERT YOUR DB NAME HERE'
host_db <- 'INSERT YOUR HOST NUMBER HERE'
db_port <- 'INSERT YOUR DB PORT HERE'
db_user <- 'INSERT YOUR USER NAME HERE'
db_password <- 'INSERT YOUR PASSWORD HERE'
con <- dbConnect(RPostgres::Postgres(), dbname = db, host=host_db, port=db_port, user=db_user, password=db_password)
dbListTables(con) | /postgre_connector.R | no_license | marcosemn/Connectors | R | false | false | 437 | r | ### Initialization ####
library('RPostgres')
library('RPostgreSQL')
### Creating connector
db <- 'INSERT YOUR DB NAME HERE'
host_db <- 'INSERT YOUR HOST NUMBER HERE'
db_port <- 'INSERT YOUR DB PORT HERE'
db_user <- 'INSERT YOUR USER NAME HERE'
db_password <- 'INSERT YOUR PASSWORD HERE'
con <- dbConnect(RPostgres::Postgres(), dbname = db, host=host_db, port=db_port, user=db_user, password=db_password)
dbListTables(con) |
# ----------------------------------------------------------------------------------------------------
# Hidden Markov Models
# Faits stylisΓ©s des rendements financiers
# ----------------------------------------------------------------------------------------------------
# written
# Gabriel LEMYRE
# ----------------------------------------------------------------------------------------------------
# Under the supervision of :
# Maciej AUGUSTYNIAK
# ----------------------------------------------------------------------------------------------------
# First version : january 8th, 2020
# Last version : february 13th, 2020
# ----------------------------------------------------------------------------------------------------
# --------------------------------------------------------
# ANALYSE D'UNE SΓRIE ET RΓSULTATS SUR LES FAITS STYLISΓS
# --------------------------------------------------------
# Fonction produisant des tables et graphiques
# L'idΓ©e est de montrer qu'une sΓ©rie prΓ©sente les
# caractΓ©ristiques des faits stylisΓ©s des rendements
# financiers
# --------------------------------------------------------
# Instalation du package forecast permettant d'utiliser les fonctions ACF, PAcf et CCf
# install.packages('forecast', dependencies = TRUE) # Installation du package
library(forecast)
# Permet de créer des graphiques très complexes
# install.packages('ggplot2', dependencies = TRUE) # Installation du package
library(ggplot2)
# Instalation du package cowplot opermettant de combiner des graphiques
# install.packages('cowplot', dependencies = TRUE) # Installation du package
library(cowplot)
# --------------------------------------------------------
# Paramètres graphiques
# --------------------------------------------------------
title.size.var <- 12
axis.lab.size.var <- 8
lty.ciline <- 2
lwd.ciline <- 0.3
lwd.predict <- 0.2
# Nombre de points Γ afficher sur l'axe horizontal
nbTicks=10
# JPEG DIMENSIONS FOR OUTPUTED FILE
image.width <- 1250
image.heigth <- 666
# Faits stylisΓ©s
title.size.var <- 12
axis.lab.size.var <- 8
resdev=300
# --------------------
# --------------------------------------------------------
# Point (1) et (2) : Absence d'acf pour rendement mais
# prΓ©sence pour erreurs^2 et abs(erreurs)
# --------------------------------------------------------
epsilon <- logR-mean(logR)
conf.level <- 0.95
ciline <- qnorm((1 - conf.level)/2)/sqrt(length(epsilon))
# Fonction d'autocorrΓ©lation des rendements
AutoCorrelation <- Acf(logR, plot = FALSE,lag.max = 200)
# # Fonction d'autocorrΓ©lation des rendements centrΓ©s au carrΓ©
AutoCorrelationCarreRES <- Acf(epsilon^2, plot = FALSE,lag.max = 200)
# # Fonction d'autocorrΓ©lation des rendements centrΓ©s en valeur absolue
AutoCorrelationAbsRES <- Acf(abs(epsilon), plot = FALSE,lag.max = 200)
A.AutoCorrelation <- with(AutoCorrelation, data.frame(lag, acf))
x <- A.AutoCorrelation$lag
y <- A.AutoCorrelation$acf
lo.A <- loess(y~x)
B.AutoCorrelation <- with(AutoCorrelationCarreRES, data.frame(lag, acf))
x <- B.AutoCorrelation$lag
y <- B.AutoCorrelation$acf
lo.B <- loess(y~x)
C.AutoCorrelation <- with(AutoCorrelationAbsRES, data.frame(lag, acf))
x <- C.AutoCorrelation$lag
y <- C.AutoCorrelation$acf
lo.C <- loess(y~x)
# DiffΓ©rence entre carrΓ© et val absolue
D.AutoCorrelation <- data.frame(B.AutoCorrelation$lag, (C.AutoCorrelation-B.AutoCorrelation)$acf)
names(D.AutoCorrelation) <- c("lag","acf")
A <- ggplot(data = A.AutoCorrelation, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title=paste("ACF des log-rendements de ",index,sep=""),
x ="Lags",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.A)), color = "red", linetype = 1,lwd=lwd.predict)
B <- ggplot(data = B.AutoCorrelation, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title=paste("ACF des log-rendements centrΓ©s au carrΓ© de ",index,sep=""),
x ="Lags",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.B)), color = "red", linetype = 1,lwd=lwd.predict)
C <- ggplot(data = C.AutoCorrelation, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title=paste("ACF de la valeur absolue des log-rendements centrΓ©s de ",index,sep=""),
x ="Lags",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.C)), color = "red", linetype = 1,lwd=lwd.predict)
D <- ggplot(data = D.AutoCorrelation, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title="DiffΓ©rence entre CorrΓ©lation valeur absolue et au carrΓ©",
x ="Lags",
y = "DiffΓ©rence") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
)
p <- plot_grid(A,B,C,D, labels = "AUTO", ncol = 1)
ggsave(paste(GraphPath,"/1 - DataStats/Dataset_ACF_",index,".png",sep=""), p)
# --------------------------------------------------------
# Point (3) : Clustering des volatilitΓ©s
# --------------------------------------------------------
# Plot the dataset and export resulting plot
jpeg(paste(GraphPath,"/1 - DataStats/Dataset_",index,".png",sep=""), width = image.width, height = image.heigth)
layout(matrix(c(1,2),2,1,byrow=TRUE))
par(mar = rep(6, 4)) # Set the margin on all sides to 2
xtick<-seq(1, length(dates), by=floor(length(dates)/nbTicks))
plot(Pt,type="l",xlab="Dates",ylab="Value", xaxt="n", cex.axis=1.8, cex.lab=2)
axis(side=1, at=xtick, labels=dates[xtick], cex.axis=1.8)
title(main=paste("Valeur de l'indice ",index,sep=""), cex.main=2)
plot(logR,type="l",xlab="Dates", xaxt="n", cex.axis=1.8, cex.lab=2)
axis(side=1, at=xtick, labels=dates[xtick], cex.axis=1.8)
title(main=paste("Log-rendement sur l'indice ",index,sep=""), cex.main=2)
abline(h=0, col="blue")
dev.off()
# --------------------------------------------------------
# Point (4) : AsymΓ©trie nΓ©gative et grand aplatissement
# --------------------------------------------------------
#descriptive statistics
des_stats <- c("Moyenne" = mean(logR),
"Γcart-type" = sd(logR),
"AsymΓ©trie" = timeDate::skewness(logR, method="moment"),
"Aplatissement" = timeDate::kurtosis(logR, method="moment"),
"Minimum" = min(logR),
"Maximum" = max(logR), "n" = length(logR)
)
des_stats
# --------------------------------------------------------
# Point (5) : Effet de levier
# --------------------------------------------------------
maxf <- function(x){max(x,0)}
negChoc <- sapply(-epsilon,FUN=maxf)
posChoc <- sapply(epsilon,FUN=maxf)
# CorrΓ©lation entre la valeur absolue des erreurs et un choc nΓ©gatif
AutoCorrelationLEVIER.NEG <- Ccf(abs(epsilon), negChoc, plot=F,lag.max = 200,level=0.95)
# CorrΓ©lation entre la valeur absolue des erreurs et un choc positif
AutoCorrelationLEVIER.POS <- Ccf(abs(epsilon), posChoc, plot=F,lag.max = 200)
NEG.AutoCorrelation <- with(AutoCorrelationLEVIER.NEG, data.frame(lag, acf))
NEG.AutoCorrelation.df <- subset(NEG.AutoCorrelation, lag >= 0)
x <- NEG.AutoCorrelation.df$lag
y <- NEG.AutoCorrelation.df$acf
lo.LEVIER.NEG <- loess(y~x)
POS.AutoCorrelation <- with(AutoCorrelationLEVIER.POS, data.frame(lag, acf))
POS.AutoCorrelation.df <- subset(POS.AutoCorrelation, lag >= 0)
x <- POS.AutoCorrelation.df$lag
y <- POS.AutoCorrelation.df$acf
lo.LEVIER.POS <- loess(y~x)
# CorrΓ©lation entre la valeur absolue des erreurs et un choc positif
AutoCorrelationLEVIER.DIFF <- data.frame(NEG.AutoCorrelation.df$lag, (NEG.AutoCorrelation.df-POS.AutoCorrelation.df)$acf)
names(AutoCorrelationLEVIER.DIFF) <- c("lag","acf")
NEG <- ggplot(data = NEG.AutoCorrelation.df, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue') +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue') +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title="CorrΓ©lation avec un choc nΓ©gatif",
x ="Lag",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.LEVIER.NEG)), color = "red", linetype = 1,lwd=lwd.predict)
POS <- ggplot(data = POS.AutoCorrelation.df, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue') +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue') +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title="CorrΓ©lation avec un choc positif",
x ="Lag",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.LEVIER.POS)), color = "red", linetype = 1,lwd=lwd.predict)
DIFF <- ggplot(data = AutoCorrelationLEVIER.DIFF, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue') +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue') +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title="CorrΓ©lation choc nΓ©gatif - CorrΓ©lation choc positif",
x ="Lag",
y = "DiffΓ©rence") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
)
p <- plot_grid(NEG,POS,DIFF, labels = "AUTO", ncol = 1)
ggsave(paste(GraphPath,"/1 - DataStats/Dataset_LEVIER_",index,".png",sep=""), p)
| /package/HMM/Faits_Stylises.R | no_license | GabrielLemyre/HHMM | R | false | false | 12,168 | r | # ----------------------------------------------------------------------------------------------------
# Hidden Markov Models
# Faits stylisΓ©s des rendements financiers
# ----------------------------------------------------------------------------------------------------
# written
# Gabriel LEMYRE
# ----------------------------------------------------------------------------------------------------
# Under the supervision of :
# Maciej AUGUSTYNIAK
# ----------------------------------------------------------------------------------------------------
# First version : january 8th, 2020
# Last version : february 13th, 2020
# ----------------------------------------------------------------------------------------------------
# --------------------------------------------------------
# ANALYSE D'UNE SΓRIE ET RΓSULTATS SUR LES FAITS STYLISΓS
# --------------------------------------------------------
# Fonction produisant des tables et graphiques
# L'idΓ©e est de montrer qu'une sΓ©rie prΓ©sente les
# caractΓ©ristiques des faits stylisΓ©s des rendements
# financiers
# --------------------------------------------------------
# Instalation du package forecast permettant d'utiliser les fonctions ACF, PAcf et CCf
# install.packages('forecast', dependencies = TRUE) # Installation du package
library(forecast)
# Permet de créer des graphiques très complexes
# install.packages('ggplot2', dependencies = TRUE) # Installation du package
library(ggplot2)
# Instalation du package cowplot opermettant de combiner des graphiques
# install.packages('cowplot', dependencies = TRUE) # Installation du package
library(cowplot)
# --------------------------------------------------------
# Paramètres graphiques
# --------------------------------------------------------
title.size.var <- 12
axis.lab.size.var <- 8
lty.ciline <- 2
lwd.ciline <- 0.3
lwd.predict <- 0.2
# Nombre de points Γ afficher sur l'axe horizontal
nbTicks=10
# JPEG DIMENSIONS FOR OUTPUTED FILE
image.width <- 1250
image.heigth <- 666
# Faits stylisΓ©s
title.size.var <- 12
axis.lab.size.var <- 8
resdev=300
# --------------------
# --------------------------------------------------------
# Point (1) et (2) : Absence d'acf pour rendement mais
# prΓ©sence pour erreurs^2 et abs(erreurs)
# --------------------------------------------------------
epsilon <- logR-mean(logR)
conf.level <- 0.95
ciline <- qnorm((1 - conf.level)/2)/sqrt(length(epsilon))
# Fonction d'autocorrΓ©lation des rendements
AutoCorrelation <- Acf(logR, plot = FALSE,lag.max = 200)
# # Fonction d'autocorrΓ©lation des rendements centrΓ©s au carrΓ©
AutoCorrelationCarreRES <- Acf(epsilon^2, plot = FALSE,lag.max = 200)
# # Fonction d'autocorrΓ©lation des rendements centrΓ©s en valeur absolue
AutoCorrelationAbsRES <- Acf(abs(epsilon), plot = FALSE,lag.max = 200)
A.AutoCorrelation <- with(AutoCorrelation, data.frame(lag, acf))
x <- A.AutoCorrelation$lag
y <- A.AutoCorrelation$acf
lo.A <- loess(y~x)
B.AutoCorrelation <- with(AutoCorrelationCarreRES, data.frame(lag, acf))
x <- B.AutoCorrelation$lag
y <- B.AutoCorrelation$acf
lo.B <- loess(y~x)
C.AutoCorrelation <- with(AutoCorrelationAbsRES, data.frame(lag, acf))
x <- C.AutoCorrelation$lag
y <- C.AutoCorrelation$acf
lo.C <- loess(y~x)
# DiffΓ©rence entre carrΓ© et val absolue
D.AutoCorrelation <- data.frame(B.AutoCorrelation$lag, (C.AutoCorrelation-B.AutoCorrelation)$acf)
names(D.AutoCorrelation) <- c("lag","acf")
A <- ggplot(data = A.AutoCorrelation, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title=paste("ACF des log-rendements de ",index,sep=""),
x ="Lags",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.A)), color = "red", linetype = 1,lwd=lwd.predict)
B <- ggplot(data = B.AutoCorrelation, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title=paste("ACF des log-rendements centrΓ©s au carrΓ© de ",index,sep=""),
x ="Lags",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.B)), color = "red", linetype = 1,lwd=lwd.predict)
C <- ggplot(data = C.AutoCorrelation, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title=paste("ACF de la valeur absolue des log-rendements centrΓ©s de ",index,sep=""),
x ="Lags",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.C)), color = "red", linetype = 1,lwd=lwd.predict)
D <- ggplot(data = D.AutoCorrelation, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue', size=lwd.ciline) +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title="DiffΓ©rence entre CorrΓ©lation valeur absolue et au carrΓ©",
x ="Lags",
y = "DiffΓ©rence") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
)
p <- plot_grid(A,B,C,D, labels = "AUTO", ncol = 1)
ggsave(paste(GraphPath,"/1 - DataStats/Dataset_ACF_",index,".png",sep=""), p)
# --------------------------------------------------------
# Point (3) : Clustering des volatilitΓ©s
# --------------------------------------------------------
# Plot the dataset and export resulting plot
jpeg(paste(GraphPath,"/1 - DataStats/Dataset_",index,".png",sep=""), width = image.width, height = image.heigth)
layout(matrix(c(1,2),2,1,byrow=TRUE))
par(mar = rep(6, 4)) # Set the margin on all sides to 2
xtick<-seq(1, length(dates), by=floor(length(dates)/nbTicks))
plot(Pt,type="l",xlab="Dates",ylab="Value", xaxt="n", cex.axis=1.8, cex.lab=2)
axis(side=1, at=xtick, labels=dates[xtick], cex.axis=1.8)
title(main=paste("Valeur de l'indice ",index,sep=""), cex.main=2)
plot(logR,type="l",xlab="Dates", xaxt="n", cex.axis=1.8, cex.lab=2)
axis(side=1, at=xtick, labels=dates[xtick], cex.axis=1.8)
title(main=paste("Log-rendement sur l'indice ",index,sep=""), cex.main=2)
abline(h=0, col="blue")
dev.off()
# --------------------------------------------------------
# Point (4) : AsymΓ©trie nΓ©gative et grand aplatissement
# --------------------------------------------------------
#descriptive statistics
des_stats <- c("Moyenne" = mean(logR),
"Γcart-type" = sd(logR),
"AsymΓ©trie" = timeDate::skewness(logR, method="moment"),
"Aplatissement" = timeDate::kurtosis(logR, method="moment"),
"Minimum" = min(logR),
"Maximum" = max(logR), "n" = length(logR)
)
des_stats
# --------------------------------------------------------
# Point (5) : Effet de levier
# --------------------------------------------------------
maxf <- function(x){max(x,0)}
negChoc <- sapply(-epsilon,FUN=maxf)
posChoc <- sapply(epsilon,FUN=maxf)
# CorrΓ©lation entre la valeur absolue des erreurs et un choc nΓ©gatif
AutoCorrelationLEVIER.NEG <- Ccf(abs(epsilon), negChoc, plot=F,lag.max = 200,level=0.95)
# CorrΓ©lation entre la valeur absolue des erreurs et un choc positif
AutoCorrelationLEVIER.POS <- Ccf(abs(epsilon), posChoc, plot=F,lag.max = 200)
NEG.AutoCorrelation <- with(AutoCorrelationLEVIER.NEG, data.frame(lag, acf))
NEG.AutoCorrelation.df <- subset(NEG.AutoCorrelation, lag >= 0)
x <- NEG.AutoCorrelation.df$lag
y <- NEG.AutoCorrelation.df$acf
lo.LEVIER.NEG <- loess(y~x)
POS.AutoCorrelation <- with(AutoCorrelationLEVIER.POS, data.frame(lag, acf))
POS.AutoCorrelation.df <- subset(POS.AutoCorrelation, lag >= 0)
x <- POS.AutoCorrelation.df$lag
y <- POS.AutoCorrelation.df$acf
lo.LEVIER.POS <- loess(y~x)
# CorrΓ©lation entre la valeur absolue des erreurs et un choc positif
AutoCorrelationLEVIER.DIFF <- data.frame(NEG.AutoCorrelation.df$lag, (NEG.AutoCorrelation.df-POS.AutoCorrelation.df)$acf)
names(AutoCorrelationLEVIER.DIFF) <- c("lag","acf")
NEG <- ggplot(data = NEG.AutoCorrelation.df, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue') +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue') +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title="CorrΓ©lation avec un choc nΓ©gatif",
x ="Lag",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.LEVIER.NEG)), color = "red", linetype = 1,lwd=lwd.predict)
POS <- ggplot(data = POS.AutoCorrelation.df, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue') +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue') +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title="CorrΓ©lation avec un choc positif",
x ="Lag",
y = "Auto-CorrΓ©lation") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
) +
geom_line(aes(x = lag, y = predict(lo.LEVIER.POS)), color = "red", linetype = 1,lwd=lwd.predict)
DIFF <- ggplot(data = AutoCorrelationLEVIER.DIFF, mapping = aes(x = lag, y = acf)) +
geom_hline(aes(yintercept = 0)) +
geom_hline(aes(yintercept = ciline), linetype = lty.ciline, color = 'darkblue') +
geom_hline(aes(yintercept = -ciline), linetype = lty.ciline, color = 'darkblue') +
geom_segment(mapping = aes(xend = lag, yend = 0)) +
labs(title="CorrΓ©lation choc nΓ©gatif - CorrΓ©lation choc positif",
x ="Lag",
y = "DiffΓ©rence") +
theme(
plot.title = element_text(color="Black", size=title.size.var, face="bold.italic"),
axis.title.x = element_text(color="black", size=axis.lab.size.var, face="bold"),
axis.title.y = element_text(color="black", size=axis.lab.size.var, face="bold")
)
p <- plot_grid(NEG,POS,DIFF, labels = "AUTO", ncol = 1)
ggsave(paste(GraphPath,"/1 - DataStats/Dataset_LEVIER_",index,".png",sep=""), p)
|
library(tidyverse)
library(lubridate)
library(janitor)
library(miceadds)
source.all("functions/")
dir.create("4_clean_data/temp")
# Home Nations
countries <- c("england", "scotland", "wales", "ireland")
for (i in 1:length(countries)){
country <- countries[i]
file_path <- paste("2_raw_data/", country, ".csv", sep = "")
country_data <- read_csv(file_path)
clean_data <- clean_rugby_data(country_data) %>%
mutate(country = str_to_title(country)) %>%
fix_home_nation_location()
dir_path <- paste("4_clean_data/", country, sep ="")
dir.create(dir_path)
clean_file_path <- paste("4_clean_data/",country, "/", country, "_clean", ".csv", sep = "")
clean_data %>%
write_csv(clean_file_path)
clean_file_path <- paste("4_clean_data/temp/", country, ".csv", sep = "")
clean_data %>%
write_csv(clean_file_path)
}
# Clean France Data
france <- read_csv("2_raw_data/france.csv")
clean_data <- clean_rugby_data(france) %>%
mutate(country = "France") %>%
fix_france_location()
dir_path <- paste("4_clean_data/france")
dir.create(dir_path)
clean_data %>%
write_csv("4_clean_data/france/france_clean.csv")
clean_data %>%
write_csv("4_clean_data/temp/france_clean.csv")
italy <- read_csv("2_raw_data/italy.csv")
clean_data <- clean_rugby_data(italy) %>%
mutate(country = "Italy") %>%
fix_italy_location()
dir_path <- paste("4_clean_data/italy")
dir.create(dir_path)
clean_data %>%
write_csv("4_clean_data/italy/italy_clean.csv")
clean_data %>%
write_csv("4_clean_data/temp/italy_clean.csv")
files <- c()
for (file in list.files("4_clean_data/temp")){
file_path <- paste("4_clean_data/temp/", file, sep = "")
files <- c(files, file_path)
}
complete_data <- NULL
for (file in files){
part_data <- read_csv(file)
complete_data <-bind_rows(complete_data, part_data)
}
complete_data %>%
relocate(country, .before = cap_no) %>%
write_csv("4_clean_data/full_table.csv")
unlink("4_clean_data/temp/", recursive = TRUE) | /3_cleaning_scripts/1_home_nations_clean_data.R | no_license | hgw2/rugby_players_project | R | false | false | 2,030 | r | library(tidyverse)
library(lubridate)
library(janitor)
library(miceadds)
source.all("functions/")
dir.create("4_clean_data/temp")
# Home Nations
countries <- c("england", "scotland", "wales", "ireland")
for (i in 1:length(countries)){
country <- countries[i]
file_path <- paste("2_raw_data/", country, ".csv", sep = "")
country_data <- read_csv(file_path)
clean_data <- clean_rugby_data(country_data) %>%
mutate(country = str_to_title(country)) %>%
fix_home_nation_location()
dir_path <- paste("4_clean_data/", country, sep ="")
dir.create(dir_path)
clean_file_path <- paste("4_clean_data/",country, "/", country, "_clean", ".csv", sep = "")
clean_data %>%
write_csv(clean_file_path)
clean_file_path <- paste("4_clean_data/temp/", country, ".csv", sep = "")
clean_data %>%
write_csv(clean_file_path)
}
# Clean France Data
france <- read_csv("2_raw_data/france.csv")
clean_data <- clean_rugby_data(france) %>%
mutate(country = "France") %>%
fix_france_location()
dir_path <- paste("4_clean_data/france")
dir.create(dir_path)
clean_data %>%
write_csv("4_clean_data/france/france_clean.csv")
clean_data %>%
write_csv("4_clean_data/temp/france_clean.csv")
italy <- read_csv("2_raw_data/italy.csv")
clean_data <- clean_rugby_data(italy) %>%
mutate(country = "Italy") %>%
fix_italy_location()
dir_path <- paste("4_clean_data/italy")
dir.create(dir_path)
clean_data %>%
write_csv("4_clean_data/italy/italy_clean.csv")
clean_data %>%
write_csv("4_clean_data/temp/italy_clean.csv")
files <- c()
for (file in list.files("4_clean_data/temp")){
file_path <- paste("4_clean_data/temp/", file, sep = "")
files <- c(files, file_path)
}
complete_data <- NULL
for (file in files){
part_data <- read_csv(file)
complete_data <-bind_rows(complete_data, part_data)
}
complete_data %>%
relocate(country, .before = cap_no) %>%
write_csv("4_clean_data/full_table.csv")
unlink("4_clean_data/temp/", recursive = TRUE) |
# Package Start-up Functions
.onAttach <- function(libname, pkgname) {
if (interactive() && is.null(options('bhklab.startup_'))) {
oldOpts <- options()
options(warn=-1)
on.exit(options(oldOpts))
packageStartupMessage(
"
RadioGx package brought to you by:
\u2588\u2588\u2588\u2588\u2588\u2588\u2557 \u2588\u2588\u2557 \u2588\u2588\u2557\u2588\u2588\u2557 \u2588\u2588\u2557\u2588\u2588\u2557 \u2588\u2588\u2588\u2588\u2588\u2557 \u2588\u2588\u2588\u2588\u2588\u2588\u2557
\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557\u2588\u2588\u2551 \u2588\u2588\u2551\u2588\u2588\u2551 \u2588\u2588\u2554\u255d\u2588\u2588\u2551 \u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557
\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2551\u2588\u2588\u2588\u2588\u2588\u2554\u255d \u2588\u2588\u2551 \u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2551\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d
\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2551\u2588\u2588\u2554\u2550\u2588\u2588\u2557 \u2588\u2588\u2551 \u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2551\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557
\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d\u2588\u2588\u2551 \u2588\u2588\u2551\u2588\u2588\u2551 \u2588\u2588\u2557\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2557\u2588\u2588\u2551 \u2588\u2588\u2551\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d
\u255a\u2550\u2550\u2550\u2550\u2550\u255d \u255a\u2550\u255d \u255a\u2550\u255d\u255a\u2550\u255d \u255a\u2550\u255d\u255a\u2550\u2550\u2550\u2550\u2550\u2550\u255d\u255a\u2550\u255d \u255a\u2550\u255d\u255a\u2550\u2550\u2550\u2550\u2550\u255d
For more of our work visit bhklab.ca!
"
)
# Prevent repeated messages when loading multiple lab packages
options(bhklab.startup_=FALSE)
}
}
| /R/zzz.R | no_license | bhklab/RadioGx | R | false | false | 1,977 | r | # Package Start-up Functions
.onAttach <- function(libname, pkgname) {
if (interactive() && is.null(options('bhklab.startup_'))) {
oldOpts <- options()
options(warn=-1)
on.exit(options(oldOpts))
packageStartupMessage(
"
RadioGx package brought to you by:
\u2588\u2588\u2588\u2588\u2588\u2588\u2557 \u2588\u2588\u2557 \u2588\u2588\u2557\u2588\u2588\u2557 \u2588\u2588\u2557\u2588\u2588\u2557 \u2588\u2588\u2588\u2588\u2588\u2557 \u2588\u2588\u2588\u2588\u2588\u2588\u2557
\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557\u2588\u2588\u2551 \u2588\u2588\u2551\u2588\u2588\u2551 \u2588\u2588\u2554\u255d\u2588\u2588\u2551 \u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557
\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2551\u2588\u2588\u2588\u2588\u2588\u2554\u255d \u2588\u2588\u2551 \u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2551\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d
\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2551\u2588\u2588\u2554\u2550\u2588\u2588\u2557 \u2588\u2588\u2551 \u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2551\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557
\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d\u2588\u2588\u2551 \u2588\u2588\u2551\u2588\u2588\u2551 \u2588\u2588\u2557\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2557\u2588\u2588\u2551 \u2588\u2588\u2551\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d
\u255a\u2550\u2550\u2550\u2550\u2550\u255d \u255a\u2550\u255d \u255a\u2550\u255d\u255a\u2550\u255d \u255a\u2550\u255d\u255a\u2550\u2550\u2550\u2550\u2550\u2550\u255d\u255a\u2550\u255d \u255a\u2550\u255d\u255a\u2550\u2550\u2550\u2550\u2550\u255d
For more of our work visit bhklab.ca!
"
)
# Prevent repeated messages when loading multiple lab packages
options(bhklab.startup_=FALSE)
}
}
|
require(dplyr)
require(ggplot2)
require(tidyr)
require(reshape2)
require(gtools)
require(mice)
require(dummies)
require(e1071)
require(mice)
require(scales)
set.seed(12345)
#load data sets
sessions_df <- read.csv("sessions.csv", header=TRUE)
train_df_start <- read.csv("train_users_2.csv", header=TRUE)
summarystats_df <- read.csv("age_gender_bkts.csv")
countries_df <- read.csv("countries.csv")
## Sessions Wrangling
sessions_df <- group_by(sessions_df, user_id)
sessions_summary <- sessions_df %>% group_by(user_id) %>% summarise(secs_elapsed_avg = mean(secs_elapsed, na.rm=TRUE), secs_elapsed_total = sum(secs_elapsed, na.rm=TRUE), secs_elapsed_sd = sd(secs_elapsed, na.rm=TRUE), secs_elapsed_min= min(secs_elapsed, na.rm=TRUE), secs_elapsed_max = max(secs_elapsed, na.rm=TRUE), secs_elapsed_median = median(secs_elapsed, na.rm=TRUE), secs_elapsed_IQR = IQR(secs_elapsed, na.rm=TRUE))
sessions_summary_pl <- sessions_df %>% group_by(user_id) %>% count(user_id)
sessions_sum2 <- select(sessions_df, user_id, device_type) %>% distinct(device_type)
colnames(sessions_summary_pl)[2] <- "total_actions"
sessions_summary <- merge(sessions_summary, sessions_summary_pl, by="user_id")
sessions_summary <- filter(sessions_summary, user_id != "")
sessions_summary[is.na(sessions_summary)==TRUE] <- 0
sessions_actions <- select(sessions_df, user_id, action)
sessions_actions2 <- melt(sessions_actions, id="user_id", na.rm=TRUE)
sessions_actions2 <- dcast(sessions_actions2, user_id ~ value + variable, length)
sessions_actions2 <- filter(sessions_actions2, user_id != "")
names(sessions_actions2)[2] <- "blank_action"
sessions_device <- select(sessions_df, user_id, device_type)
sessions_device2 <- sessions_device %>% group_by(user_id) %>% summarize(device_type = names(which.max(table(device_type))))
sessions_device2 <- filter(sessions_device2, user_id != "")
sessions_device3 <- melt(sessions_device2, id="user_id", na.rm=TRUE)
sessions_device3 <- dcast(sessions_device3, user_id ~ value + variable, length)
sessions_summary <- left_join(sessions_summary, sessions_device3, by="user_id")
sessions_final <- left_join(sessions_summary, sessions_actions2, by="user_id")
##Training set wrangling
train_df1 <- train_df_start
colnames(train_df1)[1] <- "user_id"
#Fix date objects
train_df1$date_account_created <- as.Date(train_df1$date_account_created)
train_df1$date_first_booking <- NULL
train_df1$timestamp_first_active[train_df1$timestamp_first_active == ""] <- NA
train_df1$timestamp_first_active <- as.POSIXct(as.character(train_df1$timestamp_first_active), format="%Y%m%d%H%M%S")
train_df1$timestamp_first_active <- as.Date(train_df1$timestamp_first_active, format="%Y-%m-%d")
train_df1$timestamp_first_active[is.na(train_df1$timestamp_first_active)==TRUE] <- train_df1$date_account_created
#Clean up age
train_df1$age[train_df1$age > 98] <- NA
train_df1$age[train_df1$age < 15] <- NA
train_df1$age[is.na(train_df1$age)==TRUE] <- -1
train_df1$first_affiliate_tracked[train_df1$first_affiliate_tracked == ""] <- NA
train_df1$country_destination <- as.factor(train_df1$country_destination)
#Optional NA replacement
train_df2 <- train_df1
train_df2[is.na(train_df2)] <- -1
train_dfclean <- train_df2
train_age <- train_dfclean[, c("user_id", "age")]
#train_dfclean$age <- NULL
#Feature Engineering
train_fe <- train_dfclean
train_fe <- dummy.data.frame(train_fe, names=c("gender", "signup_method", "signup_flow", "language", "affiliate_channel", "affiliate_provider", "first_affiliate_tracked", "signup_app", "first_device_type", "first_browser"), omit.constants = TRUE, sep="_", fun=as.numeric, dummy.classes = "numeric")
train_fe <- separate(train_fe, date_account_created, into=c("year_account_created", "month_account_created", "day_account_created"), sep="-")
train_fe <- separate(train_fe, timestamp_first_active, into=c("year_first_active", "month_first_active", "day_first_active"), sep="-")
train_fe <- transform(train_fe, year_account_created = as.numeric(year_account_created), month_account_created=as.numeric(month_account_created), day_account_created=as.numeric(day_account_created), year_first_active=as.numeric(year_first_active), month_first_active=as.numeric(month_first_active), day_first_active=as.numeric(day_first_active))
train_fe$destination_booked <- train_fe$country_destination != "NDF"
train_join <- train_fe
train_join[is.na(train_join)] <- -1
#Join sessions and train sets
train_full <- inner_join(train_join, sessions_final, by="user_id")
outcome_labels <- bind_cols(as.data.frame(outcome), as.data.frame(outcome.org))
outcome_labels$outcome <- as.factor(outcome_labels$outcome)
outcome_labels <- levels(outcome_labels$outcome.org)
train_boost1 <- train_full
outcome1.org <- train_boost1[, "destination_booked"]
outcome1 <- outcome1.org
num.class1 = length(levels(outcome1))
levels(outcome1) = 1:num.class1
outcome1 <- as.numeric(outcome1)
outcome.org <- train_boost1[, "country_destination"]
outcome <- outcome.org
num.class = length(levels(outcome))
levels(outcome) = 1:num.class
outcome <- as.numeric(outcome)
outcome <- outcome - 1
outcome_labels <- bind_cols(as.data.frame(outcome), as.data.frame(outcome.org))
outcome_labels$outcome <- as.factor(outcome_labels$outcome)
outcome_labels <- levels(outcome_labels$outcome.org)
train_boost1$country_destination <- NULL
train_boost1$destination_booked <- NULL
train_boost2 <- train_join
outcome2.org <- train_boost2[, "country_destination"]
outcome2 <- outcome2.org
num.class2 = length(levels(outcome2))
levels(outcome2) = 1:num.class2
outcome2 <- as.numeric(outcome2)
outcome2 <- outcome2 - 1
outcome2_labels <- bind_cols(as.data.frame(outcome2), as.data.frame(outcome2.org))
outcome2_labels$outcome <- as.factor(outcome2_labels$outcome2)
outcome2_labels <- levels(outcome2_labels$outcome2.org)
train_boost2$country_destination <- NULL
train_boost2$destination_booked <- NULL
#other junk
actions <- as.data.frame(levels(sessions_df$action))
colnames(actions) <- "actions"
#Graphs!
ggplot(train_df1, aes(x=timestamp_first_active, fill=first_affiliate_tracked)) + geom_histogram(binwidth = 250)
ggplot(train_df1, aes(x=age))+ geom_histogram()
destination_summary <- as.data.frame(summary(train_df1$country_destination))
colnames(destination_summary) <- "Count"
destination_summary$Percentages <- percent(prop.table(destination_summary$Count))
destination_summary$Destination <- rownames(destination_summary)
destination_summary <- destination_summary[,c(3,1:2)]
destination_summary <- arrange(destination_summary, desc(Count))
write.csv(destination_summary, "destination_summary.csv", row.names=TRUE)
ggplot(train_df1, aes(x=date_account_created)) + geom_histogram(color="black", fill="blue", bins=42) + ggtitle("Accounts Created Over Time") + xlab("Date Account Created") + ylab("Frequency")
agebyyear <- as.data.frame(train_df_start$age)
colnames(agebyyear) <- "Age"
agebyyear$Year <- as.numeric(train_fe$year_account_created)
ggplot(agebyyear, aes(x=Age)) + geom_histogram(breaks=c(min(agebyyear$Age), seq(10,100,5), max(agebyyear$Age)), fill="red", col="black")
ggplot(agebyyear, aes(x=Year, y=Age)) + geom_jitter(col="blue")
summary(agebyyear$Age)
summary(ageoutliers)
firstdevice <- as.data.frame(train_fe$year_account_created)
colnames(firstdevice) <- "Year"
firstdevice$First_Device <- train_df_start$first_device_type
firstdevice2010 <- filter(firstdevice, Year == 2010)
colnames(firstdevice2010) <- c("Year", "2010")
firstdevice2010table <- as.data.frame(table(firstdevice2010$"2010"))
firstdevice2010table$Freq <- prop.table(firstdevice2010table$Freq)
firstdevice2011 <- filter(firstdevice, Year == 2011)
colnames(firstdevice2011) <- c("Year", "2011")
firstdevice2011table <- as.data.frame(table(firstdevice2011$"2011"))
firstdevice2011table$Freq <- prop.table(firstdevice2011table$Freq)
firstdevice2012 <- filter(firstdevice, Year == 2012)
colnames(firstdevice2012) <- c("Year", "2012")
firstdevice2012table <- as.data.frame(table(firstdevice2012$"2012"))
firstdevice2012table$Freq <- prop.table(firstdevice2012table$Freq)
firstdevice2013 <- filter(firstdevice, Year == 2013)
colnames(firstdevice2013) <- c("Year", "2013")
firstdevice2013table <- as.data.frame(table(firstdevice2013$"2013"))
firstdevice2013table$Freq <- prop.table(firstdevice2013table$Freq)
firstdevice2014 <- filter(firstdevice, Year == 2014)
colnames(firstdevice2014) <- c("Year", "2014")
firstdevice2014table <- as.data.frame(table(firstdevice2014$"2014"))
firstdevice2014table$Freq <- prop.table(firstdevice2014table$Freq)
firstdevicefull <- full_join(firstdevice2010table, firstdevice2011table, by="Var1")
firstdevicefull <- full_join(firstdevicefull, firstdevice2012table, by="Var1")
firstdevicefull <- full_join(firstdevicefull, firstdevice2013table, by="Var1")
firstdevicefull <- full_join(firstdevicefull, firstdevice2014table, by="Var1")
colnames(firstdevicefull) <- c("Device", "2010", "2011", "2012", "2013", "2014")
firstdevicefull[,2:6] <- round(firstdevicefull[,2:6], 4)
write.csv(firstdevicefull, "firstdevicefull.csv")
| /wrangle_code.R | no_license | zrodnick/airbnb_capstone | R | false | false | 9,055 | r | require(dplyr)
require(ggplot2)
require(tidyr)
require(reshape2)
require(gtools)
require(mice)
require(dummies)
require(e1071)
require(mice)
require(scales)
set.seed(12345)
#load data sets
sessions_df <- read.csv("sessions.csv", header=TRUE)
train_df_start <- read.csv("train_users_2.csv", header=TRUE)
summarystats_df <- read.csv("age_gender_bkts.csv")
countries_df <- read.csv("countries.csv")
## Sessions Wrangling
sessions_df <- group_by(sessions_df, user_id)
sessions_summary <- sessions_df %>% group_by(user_id) %>% summarise(secs_elapsed_avg = mean(secs_elapsed, na.rm=TRUE), secs_elapsed_total = sum(secs_elapsed, na.rm=TRUE), secs_elapsed_sd = sd(secs_elapsed, na.rm=TRUE), secs_elapsed_min= min(secs_elapsed, na.rm=TRUE), secs_elapsed_max = max(secs_elapsed, na.rm=TRUE), secs_elapsed_median = median(secs_elapsed, na.rm=TRUE), secs_elapsed_IQR = IQR(secs_elapsed, na.rm=TRUE))
sessions_summary_pl <- sessions_df %>% group_by(user_id) %>% count(user_id)
sessions_sum2 <- select(sessions_df, user_id, device_type) %>% distinct(device_type)
colnames(sessions_summary_pl)[2] <- "total_actions"
sessions_summary <- merge(sessions_summary, sessions_summary_pl, by="user_id")
sessions_summary <- filter(sessions_summary, user_id != "")
sessions_summary[is.na(sessions_summary)==TRUE] <- 0
sessions_actions <- select(sessions_df, user_id, action)
sessions_actions2 <- melt(sessions_actions, id="user_id", na.rm=TRUE)
sessions_actions2 <- dcast(sessions_actions2, user_id ~ value + variable, length)
sessions_actions2 <- filter(sessions_actions2, user_id != "")
names(sessions_actions2)[2] <- "blank_action"
sessions_device <- select(sessions_df, user_id, device_type)
sessions_device2 <- sessions_device %>% group_by(user_id) %>% summarize(device_type = names(which.max(table(device_type))))
sessions_device2 <- filter(sessions_device2, user_id != "")
sessions_device3 <- melt(sessions_device2, id="user_id", na.rm=TRUE)
sessions_device3 <- dcast(sessions_device3, user_id ~ value + variable, length)
sessions_summary <- left_join(sessions_summary, sessions_device3, by="user_id")
sessions_final <- left_join(sessions_summary, sessions_actions2, by="user_id")
##Training set wrangling
train_df1 <- train_df_start
colnames(train_df1)[1] <- "user_id"
#Fix date objects
train_df1$date_account_created <- as.Date(train_df1$date_account_created)
train_df1$date_first_booking <- NULL
train_df1$timestamp_first_active[train_df1$timestamp_first_active == ""] <- NA
train_df1$timestamp_first_active <- as.POSIXct(as.character(train_df1$timestamp_first_active), format="%Y%m%d%H%M%S")
train_df1$timestamp_first_active <- as.Date(train_df1$timestamp_first_active, format="%Y-%m-%d")
train_df1$timestamp_first_active[is.na(train_df1$timestamp_first_active)==TRUE] <- train_df1$date_account_created
#Clean up age
train_df1$age[train_df1$age > 98] <- NA
train_df1$age[train_df1$age < 15] <- NA
train_df1$age[is.na(train_df1$age)==TRUE] <- -1
train_df1$first_affiliate_tracked[train_df1$first_affiliate_tracked == ""] <- NA
train_df1$country_destination <- as.factor(train_df1$country_destination)
#Optional NA replacement
train_df2 <- train_df1
train_df2[is.na(train_df2)] <- -1
train_dfclean <- train_df2
train_age <- train_dfclean[, c("user_id", "age")]
#train_dfclean$age <- NULL
#Feature Engineering
train_fe <- train_dfclean
train_fe <- dummy.data.frame(train_fe, names=c("gender", "signup_method", "signup_flow", "language", "affiliate_channel", "affiliate_provider", "first_affiliate_tracked", "signup_app", "first_device_type", "first_browser"), omit.constants = TRUE, sep="_", fun=as.numeric, dummy.classes = "numeric")
train_fe <- separate(train_fe, date_account_created, into=c("year_account_created", "month_account_created", "day_account_created"), sep="-")
train_fe <- separate(train_fe, timestamp_first_active, into=c("year_first_active", "month_first_active", "day_first_active"), sep="-")
train_fe <- transform(train_fe, year_account_created = as.numeric(year_account_created), month_account_created=as.numeric(month_account_created), day_account_created=as.numeric(day_account_created), year_first_active=as.numeric(year_first_active), month_first_active=as.numeric(month_first_active), day_first_active=as.numeric(day_first_active))
train_fe$destination_booked <- train_fe$country_destination != "NDF"
train_join <- train_fe
train_join[is.na(train_join)] <- -1
#Join sessions and train sets
train_full <- inner_join(train_join, sessions_final, by="user_id")
outcome_labels <- bind_cols(as.data.frame(outcome), as.data.frame(outcome.org))
outcome_labels$outcome <- as.factor(outcome_labels$outcome)
outcome_labels <- levels(outcome_labels$outcome.org)
train_boost1 <- train_full
outcome1.org <- train_boost1[, "destination_booked"]
outcome1 <- outcome1.org
num.class1 = length(levels(outcome1))
levels(outcome1) = 1:num.class1
outcome1 <- as.numeric(outcome1)
outcome.org <- train_boost1[, "country_destination"]
outcome <- outcome.org
num.class = length(levels(outcome))
levels(outcome) = 1:num.class
outcome <- as.numeric(outcome)
outcome <- outcome - 1
outcome_labels <- bind_cols(as.data.frame(outcome), as.data.frame(outcome.org))
outcome_labels$outcome <- as.factor(outcome_labels$outcome)
outcome_labels <- levels(outcome_labels$outcome.org)
train_boost1$country_destination <- NULL
train_boost1$destination_booked <- NULL
train_boost2 <- train_join
outcome2.org <- train_boost2[, "country_destination"]
outcome2 <- outcome2.org
num.class2 = length(levels(outcome2))
levels(outcome2) = 1:num.class2
outcome2 <- as.numeric(outcome2)
outcome2 <- outcome2 - 1
outcome2_labels <- bind_cols(as.data.frame(outcome2), as.data.frame(outcome2.org))
outcome2_labels$outcome <- as.factor(outcome2_labels$outcome2)
outcome2_labels <- levels(outcome2_labels$outcome2.org)
train_boost2$country_destination <- NULL
train_boost2$destination_booked <- NULL
#other junk
actions <- as.data.frame(levels(sessions_df$action))
colnames(actions) <- "actions"
#Graphs!
ggplot(train_df1, aes(x=timestamp_first_active, fill=first_affiliate_tracked)) + geom_histogram(binwidth = 250)
ggplot(train_df1, aes(x=age))+ geom_histogram()
destination_summary <- as.data.frame(summary(train_df1$country_destination))
colnames(destination_summary) <- "Count"
destination_summary$Percentages <- percent(prop.table(destination_summary$Count))
destination_summary$Destination <- rownames(destination_summary)
destination_summary <- destination_summary[,c(3,1:2)]
destination_summary <- arrange(destination_summary, desc(Count))
write.csv(destination_summary, "destination_summary.csv", row.names=TRUE)
ggplot(train_df1, aes(x=date_account_created)) + geom_histogram(color="black", fill="blue", bins=42) + ggtitle("Accounts Created Over Time") + xlab("Date Account Created") + ylab("Frequency")
agebyyear <- as.data.frame(train_df_start$age)
colnames(agebyyear) <- "Age"
agebyyear$Year <- as.numeric(train_fe$year_account_created)
ggplot(agebyyear, aes(x=Age)) + geom_histogram(breaks=c(min(agebyyear$Age), seq(10,100,5), max(agebyyear$Age)), fill="red", col="black")
ggplot(agebyyear, aes(x=Year, y=Age)) + geom_jitter(col="blue")
summary(agebyyear$Age)
summary(ageoutliers)
firstdevice <- as.data.frame(train_fe$year_account_created)
colnames(firstdevice) <- "Year"
firstdevice$First_Device <- train_df_start$first_device_type
firstdevice2010 <- filter(firstdevice, Year == 2010)
colnames(firstdevice2010) <- c("Year", "2010")
firstdevice2010table <- as.data.frame(table(firstdevice2010$"2010"))
firstdevice2010table$Freq <- prop.table(firstdevice2010table$Freq)
firstdevice2011 <- filter(firstdevice, Year == 2011)
colnames(firstdevice2011) <- c("Year", "2011")
firstdevice2011table <- as.data.frame(table(firstdevice2011$"2011"))
firstdevice2011table$Freq <- prop.table(firstdevice2011table$Freq)
firstdevice2012 <- filter(firstdevice, Year == 2012)
colnames(firstdevice2012) <- c("Year", "2012")
firstdevice2012table <- as.data.frame(table(firstdevice2012$"2012"))
firstdevice2012table$Freq <- prop.table(firstdevice2012table$Freq)
firstdevice2013 <- filter(firstdevice, Year == 2013)
colnames(firstdevice2013) <- c("Year", "2013")
firstdevice2013table <- as.data.frame(table(firstdevice2013$"2013"))
firstdevice2013table$Freq <- prop.table(firstdevice2013table$Freq)
firstdevice2014 <- filter(firstdevice, Year == 2014)
colnames(firstdevice2014) <- c("Year", "2014")
firstdevice2014table <- as.data.frame(table(firstdevice2014$"2014"))
firstdevice2014table$Freq <- prop.table(firstdevice2014table$Freq)
firstdevicefull <- full_join(firstdevice2010table, firstdevice2011table, by="Var1")
firstdevicefull <- full_join(firstdevicefull, firstdevice2012table, by="Var1")
firstdevicefull <- full_join(firstdevicefull, firstdevice2013table, by="Var1")
firstdevicefull <- full_join(firstdevicefull, firstdevice2014table, by="Var1")
colnames(firstdevicefull) <- c("Device", "2010", "2011", "2012", "2013", "2014")
firstdevicefull[,2:6] <- round(firstdevicefull[,2:6], 4)
write.csv(firstdevicefull, "firstdevicefull.csv")
|
library(xml2)
library(rvest)
setwd("C:/Users/xia/Desktop/splcml-result/data1/")
datasets <-c('nr','gpcr','ic','e')
for (j in 1:4){
data <-read.csv(paste0(datasets[j],"_admat_dgc.csv"),header=F)
data<- t(data[1,-1])
m<-dim(data)[1]
dgname <- matrix(data=0,nrow = m,ncol = 1)
for (i in 1:m){
site <- paste0("https://www.genome.jp/dbget-bin/www_bget?dr:",data[i,1])
webpage <- read_html(site)
name1 <- html_nodes(webpage,'.td51 div div ')
name1 <- gsub("<div style=\"width:555px;overflow-x:auto;overflow-y:hidden\">","",name1)
name1 <- gsub("<br>\n</div>","",name1)
name1 <- unlist(strsplit(name1, ";"))
name1 <- name1[1]
dgname[i] <- name1
}
drugname <-data.frame(data,dgname)
colnames(drugname) <- c('drugID','drugName')
write.csv(drugname,paste0("drugName_",datasets[j],".csv"),row.names = F)
} | /code for SPLCMF/rcode/drugIDtoName.R | no_license | Macau-LYXia/SPLCMF-DTI | R | false | false | 846 | r | library(xml2)
library(rvest)
setwd("C:/Users/xia/Desktop/splcml-result/data1/")
datasets <-c('nr','gpcr','ic','e')
for (j in 1:4){
data <-read.csv(paste0(datasets[j],"_admat_dgc.csv"),header=F)
data<- t(data[1,-1])
m<-dim(data)[1]
dgname <- matrix(data=0,nrow = m,ncol = 1)
for (i in 1:m){
site <- paste0("https://www.genome.jp/dbget-bin/www_bget?dr:",data[i,1])
webpage <- read_html(site)
name1 <- html_nodes(webpage,'.td51 div div ')
name1 <- gsub("<div style=\"width:555px;overflow-x:auto;overflow-y:hidden\">","",name1)
name1 <- gsub("<br>\n</div>","",name1)
name1 <- unlist(strsplit(name1, ";"))
name1 <- name1[1]
dgname[i] <- name1
}
drugname <-data.frame(data,dgname)
colnames(drugname) <- c('drugID','drugName')
write.csv(drugname,paste0("drugName_",datasets[j],".csv"),row.names = F)
} |
#how much memory the dataset will require
print(object.size(read.delim("hh_power.txt", sep = ";",stringsAsFactors = FALSE)),units="Mb")
#its okay to load the whole file & take a subset after
power <- read.delim("hh_power.txt", sep = ";",stringsAsFactors = FALSE)
#subset data
powerss <- power[power$Date=="1/2/2007" | power$Date=="2/2/2007",]
#formatting the date-variable
d <- paste0(0, substr(powerss$Date,1,1))
m <- paste0(0, substr(powerss$Date,3,3))
y <- substr(powerss$Date,5,8)
powerss$Date <- strptime(paste0(d,"/",m,"/",y),"%d/%m/%Y")
powerss$Date <- as.Date(powerss$Date,"%Y-%m-%d")
#powerss$Time <- strptime(powerss$Time, "%T")
powerss$Global_active_power <- as.numeric(powerss$Global_active_power)
powerss$Global_reactive_power <- as.numeric(powerss$Global_reactive_power)
powerss$Voltage <- as.numeric(powerss$Voltage)
powerss$Global_intensity <- as.numeric(powerss$Global_intensity)
powerss$Sub_metering_1 <- as.numeric(powerss$Sub_metering_1)
powerss$Sub_metering_2 <- as.numeric(powerss$Sub_metering_2)
powerss$Sub_metering_3 <- as.numeric(powerss$Sub_metering_3)
#Plot 4
png(filename = "plot4.png",
width = 480, height = 480, units = "px", pointsize = 12,
bg = "white")
par(mfrow=c(2,2))
#1
plot(powerss$Global_active_power, type="l", ylab="Global Active Power (kilowatts)", xlab="", xaxt="n")
axis(1, at=c(0,middle,right),labels=c("Thu","Fri","Sat"), las=0)
#2
plot(powerss$Voltage, type="l", ylab="Voltage", xlab="datetime", xaxt="n")
axis(1, at=c(0,middle,right),labels=c("Thu","Fri","Sat"), las=0)
#3
plot(powerss$Sub_metering_1, type="l", ylab="Energy sub metering", xlab="", xaxt="n")
axis(1, at=c(0,middle,right),labels=c("Thu","Fri","Sat"), las=0)
lines(powerss$Sub_metering_2, col="red")
lines(powerss$Sub_metering_3, col="blue")
legend("topright", legend=c("Sub_metering_1","Sub_metering_2","Sub_metering_3"),cex=0.5, lty=c(1,1,1), col=c("black","red", "blue"))
#4
plot(powerss$Global_reactive_power, type="l", ylab="Global_reactive_power", xlab="datetime", xaxt="n")
axis(1, at=c(0,middle,right),labels=c("Thu","Fri","Sat"), las=0)
par(mfrow=c(1,1))
dev.off() | /plot4.R | no_license | mamack/exploratory-graphs | R | false | false | 2,096 | r | #how much memory the dataset will require
print(object.size(read.delim("hh_power.txt", sep = ";",stringsAsFactors = FALSE)),units="Mb")
#its okay to load the whole file & take a subset after
power <- read.delim("hh_power.txt", sep = ";",stringsAsFactors = FALSE)
#subset data
powerss <- power[power$Date=="1/2/2007" | power$Date=="2/2/2007",]
#formatting the date-variable
d <- paste0(0, substr(powerss$Date,1,1))
m <- paste0(0, substr(powerss$Date,3,3))
y <- substr(powerss$Date,5,8)
powerss$Date <- strptime(paste0(d,"/",m,"/",y),"%d/%m/%Y")
powerss$Date <- as.Date(powerss$Date,"%Y-%m-%d")
#powerss$Time <- strptime(powerss$Time, "%T")
powerss$Global_active_power <- as.numeric(powerss$Global_active_power)
powerss$Global_reactive_power <- as.numeric(powerss$Global_reactive_power)
powerss$Voltage <- as.numeric(powerss$Voltage)
powerss$Global_intensity <- as.numeric(powerss$Global_intensity)
powerss$Sub_metering_1 <- as.numeric(powerss$Sub_metering_1)
powerss$Sub_metering_2 <- as.numeric(powerss$Sub_metering_2)
powerss$Sub_metering_3 <- as.numeric(powerss$Sub_metering_3)
#Plot 4
png(filename = "plot4.png",
width = 480, height = 480, units = "px", pointsize = 12,
bg = "white")
par(mfrow=c(2,2))
#1
plot(powerss$Global_active_power, type="l", ylab="Global Active Power (kilowatts)", xlab="", xaxt="n")
axis(1, at=c(0,middle,right),labels=c("Thu","Fri","Sat"), las=0)
#2
plot(powerss$Voltage, type="l", ylab="Voltage", xlab="datetime", xaxt="n")
axis(1, at=c(0,middle,right),labels=c("Thu","Fri","Sat"), las=0)
#3
plot(powerss$Sub_metering_1, type="l", ylab="Energy sub metering", xlab="", xaxt="n")
axis(1, at=c(0,middle,right),labels=c("Thu","Fri","Sat"), las=0)
lines(powerss$Sub_metering_2, col="red")
lines(powerss$Sub_metering_3, col="blue")
legend("topright", legend=c("Sub_metering_1","Sub_metering_2","Sub_metering_3"),cex=0.5, lty=c(1,1,1), col=c("black","red", "blue"))
#4
plot(powerss$Global_reactive_power, type="l", ylab="Global_reactive_power", xlab="datetime", xaxt="n")
axis(1, at=c(0,middle,right),labels=c("Thu","Fri","Sat"), las=0)
par(mfrow=c(1,1))
dev.off() |
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/nma.R
\name{nameTreatments}
\alias{nameTreatments}
\title{Match treatment names to ID numbers}
\usage{
nameTreatments(results, coding, ...)
}
\arguments{
\item{results}{A data frame as returned by \code{extractComparison}}
\item{coding}{A data frame with two columns 'id' and 'description'. 'id' must
be the treatment id numbers corresponding to the way the treatments were
coded in the network. 'description' should be the name of the treatment.}
}
\value{
The same data frame with the treatment names appended
}
\description{
Match treatment names to ID numbers
}
\details{
This function matches the coded treatment id numbers in the network
to the corresponding human readable names. The mapping from id number to
name should be provided as a two column data frame via the \code{coding}
argument. This function is intended to work on the data frame generated as
the output from \code{extractComparison}. This function will mainly be
called via \code{extractMTCResults} and should only be used directly if you
understand what you are doing. The general work flow is \code{mtc.run} >
\code{calcAllPairs} > \code{extractComparison} > \code{nameTreatments} >
\code{makeTab}. \code{extractMTCResults} will handle the last four steps
for you.
}
\seealso{
\code{\link{extractComparison}}, \code{\link{calcAllPairs}}, \code{\link{extractMTCResults}}
}
| /man/nameTreatments.Rd | no_license | RichardBirnie/mautils | R | false | true | 1,444 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/nma.R
\name{nameTreatments}
\alias{nameTreatments}
\title{Match treatment names to ID numbers}
\usage{
nameTreatments(results, coding, ...)
}
\arguments{
\item{results}{A data frame as returned by \code{extractComparison}}
\item{coding}{A data frame with two columns 'id' and 'description'. 'id' must
be the treatment id numbers corresponding to the way the treatments were
coded in the network. 'description' should be the name of the treatment.}
}
\value{
The same data frame with the treatment names appended
}
\description{
Match treatment names to ID numbers
}
\details{
This function matches the coded treatment id numbers in the network
to the corresponding human readable names. The mapping from id number to
name should be provided as a two column data frame via the \code{coding}
argument. This function is intended to work on the data frame generated as
the output from \code{extractComparison}. This function will mainly be
called via \code{extractMTCResults} and should only be used directly if you
understand what you are doing. The general work flow is \code{mtc.run} >
\code{calcAllPairs} > \code{extractComparison} > \code{nameTreatments} >
\code{makeTab}. \code{extractMTCResults} will handle the last four steps
for you.
}
\seealso{
\code{\link{extractComparison}}, \code{\link{calcAllPairs}}, \code{\link{extractMTCResults}}
}
|
#' @param taxonKey A taxon key from the GBIF backbone. All included and synonym taxa
#' are included in the search, so a search for aves with taxononKey=212
#' (i.e. /occurrence/search?taxonKey=212) will match all birds, no matter which
#' species. You can pass many keys by passing occ_search in a call to an
#' lapply-family function (see last example below).
#' @param scientificName A scientific name from the GBIF backbone. All included and synonym
#' taxa are included in the search.
#' @param datasetKey The occurrence dataset key (a uuid)
#' @param catalogNumber An identifier of any form assigned by the source within a
#' physical collection or digital dataset for the record which may not unique,
#' but should be fairly unique in combination with the institution and collection code.
#' @param collectorName The person who recorded the occurrence.
#' @param collectionCode An identifier of any form assigned by the source to identify
#' the physical collection or digital dataset uniquely within the text of an institution.
#' @param institutionCode An identifier of any form assigned by the source to identify
#' the institution the record belongs to. Not guaranteed to be que.
#' @param country The 2-letter country code (as per ISO-3166-1) of the country in
#' which the occurrence was recorded. See here
#' \url{http://en.wikipedia.org/wiki/ISO_3166-1_alpha-2}
#' @param basisOfRecord Basis of record, as defined in our BasisOfRecord enum here
#' \url{http://bit.ly/19kBGhG}. Acceptable values are:
#' \itemize{
#' \item FOSSIL_SPECIMEN An occurrence record describing a fossilized specimen.
#' \item HUMAN_OBSERVATION An occurrence record describing an observation made by
#' one or more people.
#' \item LITERATURE An occurrence record based on literature alone.
#' \item LIVING_SPECIMEN An occurrence record describing a living specimen, e.g.
#' \item MACHINE_OBSERVATION An occurrence record describing an observation made
#' by a machine.
#' \item OBSERVATION An occurrence record describing an observation.
#' \item PRESERVED_SPECIMEN An occurrence record describing a preserved specimen.
#' \item UNKNOWN Unknown basis for the record.
#' }
#' @param eventDate Occurrence date in ISO 8601 format: yyyy, yyyy-MM, yyyy-MM-dd, or
#' MM-dd.
#' @param year The 4 digit year. A year of 98 will be interpreted as AD 98.
#' @param month The month of the year, starting with 1 for January.
#' @param search Query terms. The value for this parameter can be a simple word or a phrase.
#' @param decimalLatitude Latitude in decimals between -90 and 90 based on WGS 84.
#' Supports range queries.
#' @param decimalLongitude Longitude in decimals between -180 and 180 based on WGS 84.
#' Supports range queries.
#' @param publishingCountry The 2-letter country code (as per ISO-3166-1) of the
#' country in which the occurrence was recorded.
#' @param elevation Elevation in meters above sea level.
#' @param depth Depth in meters relative to elevation. For example 10 meters below a
#' lake surface with given elevation.
#' @param geometry Searches for occurrences inside a polygon described in Well Known
#' Text (WKT) format. A WKT shape written as either POINT, LINESTRING, LINEARRING
#' or POLYGON. Example of a polygon: ((30.1 10.1, 20, 20 40, 40 40, 30.1 10.1))
#' would be queried as \url{http://bit.ly/HwUSif}.
#' @param spatialIssues (logical) Includes/excludes occurrence records which contain spatial
#' issues (as determined in our record interpretation), i.e. spatialIssues=TRUE
#' returns only those records with spatial issues while spatialIssues=FALSE includes
#' only records without spatial issues. The absence of this parameter returns any
#' record with or without spatial issues.
#' @param issue (character) One of many possible issues with each occurrence record. See
#' Details.
#' @param hasCoordinate (logical) Return only occurence records with lat/long data (TRUE) or
#' all records (FALSE, default).
#' @param typeStatus Type status of the specimen. One of many options. See ?typestatus
#' @param recordNumber Number recorded by collector of the data, different from GBIF record
#' number. See \url{http://rs.tdwg.org/dwc/terms/#recordNumber} for more info
#' @param lastInterpreted Date the record was last modified in GBIF, in ISO 8601 format:
#' yyyy, yyyy-MM, yyyy-MM-dd, or MM-dd. Supports range queries.
#' @param continent Continent. One of africa, antarctica, asia, europe, north_america
#' (North America includes the Caribbean and reachies down and includes Panama), oceania,
#' or south_america
#' @param fields (character) Default ('minimal') will return just taxon name, key, latitude, and
#' longitute. 'all' returns all fields. Or specify each field you want returned by name, e.g.
#' fields = c('name','latitude','elevation').
#' @param return One of data, hier, meta, or all. If data, a data.frame with the
#' data. hier returns the classifications in a list for each record. meta
#' returns the metadata for the entire call. all gives all data back in a list.
#' @param mediatype Media type. Default is NULL, so no filtering on mediatype. Options:
#' NULL, 'MovingImage', 'Sound', and 'StillImage'.``
#' @return A data.frame or list
#' @description
#' Note that you can pass in a vector to one of taxonkey, datasetKey, and
#' catalogNumber parameters in a function call, but not a vector >1 of the three
#' parameters at the same time
#'
#' \bold{Hierarchies:} hierarchies are returned wih each occurrence object. There is no
#' option no to return them from the API. However, within the \code{occ_search}
#' function you can select whether to return just hierarchies, just data, all of
#' data and hiearchies and metadata, or just metadata. If all hierarchies are the
#' same we just return one for you.
#'
#' \bold{Data:} By default only three data fields are returned: name (the species name),
#' decimallatitude, and decimallongitude. Set parameter minimal=FALSE if you want more data.
#'
#' \bold{Nerds:} You can pass parameters not defined in this function into the call to
#' the GBIF API to control things about the call itself using the \code{callopts}
#' function. See an example below that passes in the \code{verbose} function to
#' get details on the http call.
#'
#' \bold{Scientific names vs. taxon keys:} In the previous GBIF API and the version of rgbif that wrapped
#' that API, you could search the equivalent of this function with a species name, which was
#' convenient. However, names are messy right. So it sorta makes sense to sort out the species
#' key numbers you want exactly, and then get your occurrence data with this function. GBIF has
#' added a parameter scientificName to allow searches by scientific names in this function - which
#' includes synonym taxa.
#'
#' \bold{WKT:} Examples of valid WKT objects:
#' \itemize{
#' \item 'POLYGON((30.1 10.1, 10 20, 20 60, 60 60, 30.1 10.1))'
#' \item 'POINT(30.1 10.1)'
#' \item 'LINESTRING(3 4,10 50,20 25)'
#' \item 'LINEARRING' ???' - Not sure how to specify this. Anyone?
#' }
#'
#' \bold{Range queries:} A range query is as it sounds - you query on a range of values defined by
#' a lower and upper limit. Do a range query by specifying the lower and upper limit in a vector
#' like \code{depth='50,100'}. It would be more R like to specify the range in a vector like
#' \code{c(50,100)}, but that sort of syntax allows you to do many searches, one for each element in
#' the vector - thus range queries have to differ. The following parameters support range queries.
#' \itemize{
#' \item decimalLatitude
#' \item decimalLongitude
#' \item depth
#' \item elevation
#' \item eventDate
#' \item lastInterpreted
#' \item month
#' \item year
#' }
#'
#' \bold{Issue:} The options for the issue parameter (from
#' http://gbif.github.io/gbif-api/apidocs/org/gbif/api/vocabulary/OccurrenceIssue.html):
#' \itemize{
#' \item BASIS_OF_RECORD_INVALID The given basis of record is impossible to interpret or seriously
#' different from the recommended vocabulary.
#' \item CONTINENT_COUNTRY_MISMATCH The interpreted continent and country do not match up.
#' \item CONTINENT_DERIVED_FROM_COORDINATES The interpreted continent is based on the coordinates,
#' not the verbatim string information.
#' \item CONTINENT_INVALID Uninterpretable continent values found.
#' \item COORDINATE_INVALID Coordinate value given in some form but GBIF is unable to interpret it.
#' \item COORDINATE_OUT_OF_RANGE Coordinate has invalid lat/lon values out of their decimal max
#' range.
#' \item COORDINATE_REPROJECTED The original coordinate was successfully reprojected from a
#' different geodetic datum to WGS84.
#' \item COORDINATE_REPROJECTION_FAILED The given decimal latitude and longitude could not be
#' reprojected to WGS84 based on the provided datum.
#' \item COORDINATE_REPROJECTION_SUSPICIOUS Indicates successful coordinate reprojection according
#' to provided datum, but which results in a datum shift larger than 0.1 decimal degrees.
#' \item COORDINATE_ROUNDED Original coordinate modified by rounding to 5 decimals.
#' \item COUNTRY_COORDINATE_MISMATCH The interpreted occurrence coordinates fall outside of the
#' indicated country.
#' \item COUNTRY_DERIVED_FROM_COORDINATES The interpreted country is based on the coordinates, not
#' the verbatim string information.
#' \item COUNTRY_INVALID Uninterpretable country values found.
#' \item COUNTRY_MISMATCH Interpreted country for dwc:country and dwc:countryCode contradict each
#' other.
#' \item DEPTH_MIN_MAX_SWAPPED Set if supplied min>max
#' \item DEPTH_NON_NUMERIC Set if depth is a non numeric value
#' \item DEPTH_NOT_METRIC Set if supplied depth is not given in the metric system, for example
#' using feet instead of meters
#' \item DEPTH_UNLIKELY Set if depth is larger than 11.000m or negative.
#' \item ELEVATION_MIN_MAX_SWAPPED Set if supplied min > max elevation
#' \item ELEVATION_NON_NUMERIC Set if elevation is a non numeric value
#' \item ELEVATION_NOT_METRIC Set if supplied elevation is not given in the metric system, for
#' example using feet instead of meters
#' \item ELEVATION_UNLIKELY Set if elevation is above the troposphere (17km) or below 11km
#' (Mariana Trench).
#' \item GEODETIC_DATUM_ASSUMED_WGS84 Indicating that the interpreted coordinates assume they are
#' based on WGS84 datum as the datum was either not indicated or interpretable.
#' \item GEODETIC_DATUM_INVALID The geodetic datum given could not be interpreted.
#' \item IDENTIFIED_DATE_INVALID The date given for dwc:dateIdentified is invalid and cant be
#' interpreted at all.
#' \item IDENTIFIED_DATE_UNLIKELY The date given for dwc:dateIdentified is in the future or before
#' Linnean times (1700).
#' \item MODIFIED_DATE_INVALID A (partial) invalid date is given for dc:modified, such as a non
#' existing date, invalid zero month, etc.
#' \item MODIFIED_DATE_UNLIKELY The date given for dc:modified is in the future or predates unix
#' time (1970).
#' \item MULTIMEDIA_DATE_INVALID An invalid date is given for dc:created of a multimedia object.
#' \item MULTIMEDIA_URI_INVALID An invalid uri is given for a multimedia object.
#' \item PRESUMED_NEGATED_LATITUDE Latitude appears to be negated, e.g.
#' \item PRESUMED_NEGATED_LONGITUDE Longitude appears to be negated, e.g.
#' \item PRESUMED_SWAPPED_COORDINATE Latitude and longitude appear to be swapped.
#' \item RECORDED_DATE_INVALID A (partial) invalid date is given, such as a non existing date,
#' invalid zero month, etc.
#' \item RECORDED_DATE_MISMATCH The recording date specified as the eventDate string and the
#' individual year, month, day are contradicting.
#' \item RECORDED_DATE_UNLIKELY The recording date is highly unlikely, falling either into the
#' future or represents a very old date before 1600 that predates modern taxonomy.
#' \item REFERENCES_URI_INVALID An invalid uri is given for dc:references.
#' \item TAXON_MATCH_FUZZY Matching to the taxonomic backbone can only be done using a fuzzy, non
#' exact match.
#' \item TAXON_MATCH_HIGHERRANK Matching to the taxonomic backbone can only be done on a higher
#' rank and not the scientific name.
#' \item TAXON_MATCH_NONE Matching to the taxonomic backbone cannot be done cause there was no
#' match at all or several matches with too little information to keep them apart (homonyms).
#' \item TYPE_STATUS_INVALID The given type status is impossible to interpret or seriously
#' different from the recommended vocabulary.
#' \item ZERO_COORDINATE Coordinate is the exact 0/0 coordinate, often indicating a bad null
#' coordinate.
#' }
| /man-roxygen/occsearch.r | permissive | jarioksa/rgbif | R | false | false | 12,780 | r | #' @param taxonKey A taxon key from the GBIF backbone. All included and synonym taxa
#' are included in the search, so a search for aves with taxononKey=212
#' (i.e. /occurrence/search?taxonKey=212) will match all birds, no matter which
#' species. You can pass many keys by passing occ_search in a call to an
#' lapply-family function (see last example below).
#' @param scientificName A scientific name from the GBIF backbone. All included and synonym
#' taxa are included in the search.
#' @param datasetKey The occurrence dataset key (a uuid)
#' @param catalogNumber An identifier of any form assigned by the source within a
#' physical collection or digital dataset for the record which may not unique,
#' but should be fairly unique in combination with the institution and collection code.
#' @param collectorName The person who recorded the occurrence.
#' @param collectionCode An identifier of any form assigned by the source to identify
#' the physical collection or digital dataset uniquely within the text of an institution.
#' @param institutionCode An identifier of any form assigned by the source to identify
#' the institution the record belongs to. Not guaranteed to be que.
#' @param country The 2-letter country code (as per ISO-3166-1) of the country in
#' which the occurrence was recorded. See here
#' \url{http://en.wikipedia.org/wiki/ISO_3166-1_alpha-2}
#' @param basisOfRecord Basis of record, as defined in our BasisOfRecord enum here
#' \url{http://bit.ly/19kBGhG}. Acceptable values are:
#' \itemize{
#' \item FOSSIL_SPECIMEN An occurrence record describing a fossilized specimen.
#' \item HUMAN_OBSERVATION An occurrence record describing an observation made by
#' one or more people.
#' \item LITERATURE An occurrence record based on literature alone.
#' \item LIVING_SPECIMEN An occurrence record describing a living specimen, e.g.
#' \item MACHINE_OBSERVATION An occurrence record describing an observation made
#' by a machine.
#' \item OBSERVATION An occurrence record describing an observation.
#' \item PRESERVED_SPECIMEN An occurrence record describing a preserved specimen.
#' \item UNKNOWN Unknown basis for the record.
#' }
#' @param eventDate Occurrence date in ISO 8601 format: yyyy, yyyy-MM, yyyy-MM-dd, or
#' MM-dd.
#' @param year The 4 digit year. A year of 98 will be interpreted as AD 98.
#' @param month The month of the year, starting with 1 for January.
#' @param search Query terms. The value for this parameter can be a simple word or a phrase.
#' @param decimalLatitude Latitude in decimals between -90 and 90 based on WGS 84.
#' Supports range queries.
#' @param decimalLongitude Longitude in decimals between -180 and 180 based on WGS 84.
#' Supports range queries.
#' @param publishingCountry The 2-letter country code (as per ISO-3166-1) of the
#' country in which the occurrence was recorded.
#' @param elevation Elevation in meters above sea level.
#' @param depth Depth in meters relative to elevation. For example 10 meters below a
#' lake surface with given elevation.
#' @param geometry Searches for occurrences inside a polygon described in Well Known
#' Text (WKT) format. A WKT shape written as either POINT, LINESTRING, LINEARRING
#' or POLYGON. Example of a polygon: ((30.1 10.1, 20, 20 40, 40 40, 30.1 10.1))
#' would be queried as \url{http://bit.ly/HwUSif}.
#' @param spatialIssues (logical) Includes/excludes occurrence records which contain spatial
#' issues (as determined in our record interpretation), i.e. spatialIssues=TRUE
#' returns only those records with spatial issues while spatialIssues=FALSE includes
#' only records without spatial issues. The absence of this parameter returns any
#' record with or without spatial issues.
#' @param issue (character) One of many possible issues with each occurrence record. See
#' Details.
#' @param hasCoordinate (logical) Return only occurence records with lat/long data (TRUE) or
#' all records (FALSE, default).
#' @param typeStatus Type status of the specimen. One of many options. See ?typestatus
#' @param recordNumber Number recorded by collector of the data, different from GBIF record
#' number. See \url{http://rs.tdwg.org/dwc/terms/#recordNumber} for more info
#' @param lastInterpreted Date the record was last modified in GBIF, in ISO 8601 format:
#' yyyy, yyyy-MM, yyyy-MM-dd, or MM-dd. Supports range queries.
#' @param continent Continent. One of africa, antarctica, asia, europe, north_america
#' (North America includes the Caribbean and reachies down and includes Panama), oceania,
#' or south_america
#' @param fields (character) Default ('minimal') will return just taxon name, key, latitude, and
#' longitute. 'all' returns all fields. Or specify each field you want returned by name, e.g.
#' fields = c('name','latitude','elevation').
#' @param return One of data, hier, meta, or all. If data, a data.frame with the
#' data. hier returns the classifications in a list for each record. meta
#' returns the metadata for the entire call. all gives all data back in a list.
#' @param mediatype Media type. Default is NULL, so no filtering on mediatype. Options:
#' NULL, 'MovingImage', 'Sound', and 'StillImage'.``
#' @return A data.frame or list
#' @description
#' Note that you can pass in a vector to one of taxonkey, datasetKey, and
#' catalogNumber parameters in a function call, but not a vector >1 of the three
#' parameters at the same time
#'
#' \bold{Hierarchies:} hierarchies are returned wih each occurrence object. There is no
#' option no to return them from the API. However, within the \code{occ_search}
#' function you can select whether to return just hierarchies, just data, all of
#' data and hiearchies and metadata, or just metadata. If all hierarchies are the
#' same we just return one for you.
#'
#' \bold{Data:} By default only three data fields are returned: name (the species name),
#' decimallatitude, and decimallongitude. Set parameter minimal=FALSE if you want more data.
#'
#' \bold{Nerds:} You can pass parameters not defined in this function into the call to
#' the GBIF API to control things about the call itself using the \code{callopts}
#' function. See an example below that passes in the \code{verbose} function to
#' get details on the http call.
#'
#' \bold{Scientific names vs. taxon keys:} In the previous GBIF API and the version of rgbif that wrapped
#' that API, you could search the equivalent of this function with a species name, which was
#' convenient. However, names are messy right. So it sorta makes sense to sort out the species
#' key numbers you want exactly, and then get your occurrence data with this function. GBIF has
#' added a parameter scientificName to allow searches by scientific names in this function - which
#' includes synonym taxa.
#'
#' \bold{WKT:} Examples of valid WKT objects:
#' \itemize{
#' \item 'POLYGON((30.1 10.1, 10 20, 20 60, 60 60, 30.1 10.1))'
#' \item 'POINT(30.1 10.1)'
#' \item 'LINESTRING(3 4,10 50,20 25)'
#' \item 'LINEARRING' ???' - Not sure how to specify this. Anyone?
#' }
#'
#' \bold{Range queries:} A range query is as it sounds - you query on a range of values defined by
#' a lower and upper limit. Do a range query by specifying the lower and upper limit in a vector
#' like \code{depth='50,100'}. It would be more R like to specify the range in a vector like
#' \code{c(50,100)}, but that sort of syntax allows you to do many searches, one for each element in
#' the vector - thus range queries have to differ. The following parameters support range queries.
#' \itemize{
#' \item decimalLatitude
#' \item decimalLongitude
#' \item depth
#' \item elevation
#' \item eventDate
#' \item lastInterpreted
#' \item month
#' \item year
#' }
#'
#' \bold{Issue:} The options for the issue parameter (from
#' http://gbif.github.io/gbif-api/apidocs/org/gbif/api/vocabulary/OccurrenceIssue.html):
#' \itemize{
#' \item BASIS_OF_RECORD_INVALID The given basis of record is impossible to interpret or seriously
#' different from the recommended vocabulary.
#' \item CONTINENT_COUNTRY_MISMATCH The interpreted continent and country do not match up.
#' \item CONTINENT_DERIVED_FROM_COORDINATES The interpreted continent is based on the coordinates,
#' not the verbatim string information.
#' \item CONTINENT_INVALID Uninterpretable continent values found.
#' \item COORDINATE_INVALID Coordinate value given in some form but GBIF is unable to interpret it.
#' \item COORDINATE_OUT_OF_RANGE Coordinate has invalid lat/lon values out of their decimal max
#' range.
#' \item COORDINATE_REPROJECTED The original coordinate was successfully reprojected from a
#' different geodetic datum to WGS84.
#' \item COORDINATE_REPROJECTION_FAILED The given decimal latitude and longitude could not be
#' reprojected to WGS84 based on the provided datum.
#' \item COORDINATE_REPROJECTION_SUSPICIOUS Indicates successful coordinate reprojection according
#' to provided datum, but which results in a datum shift larger than 0.1 decimal degrees.
#' \item COORDINATE_ROUNDED Original coordinate modified by rounding to 5 decimals.
#' \item COUNTRY_COORDINATE_MISMATCH The interpreted occurrence coordinates fall outside of the
#' indicated country.
#' \item COUNTRY_DERIVED_FROM_COORDINATES The interpreted country is based on the coordinates, not
#' the verbatim string information.
#' \item COUNTRY_INVALID Uninterpretable country values found.
#' \item COUNTRY_MISMATCH Interpreted country for dwc:country and dwc:countryCode contradict each
#' other.
#' \item DEPTH_MIN_MAX_SWAPPED Set if supplied min>max
#' \item DEPTH_NON_NUMERIC Set if depth is a non numeric value
#' \item DEPTH_NOT_METRIC Set if supplied depth is not given in the metric system, for example
#' using feet instead of meters
#' \item DEPTH_UNLIKELY Set if depth is larger than 11.000m or negative.
#' \item ELEVATION_MIN_MAX_SWAPPED Set if supplied min > max elevation
#' \item ELEVATION_NON_NUMERIC Set if elevation is a non numeric value
#' \item ELEVATION_NOT_METRIC Set if supplied elevation is not given in the metric system, for
#' example using feet instead of meters
#' \item ELEVATION_UNLIKELY Set if elevation is above the troposphere (17km) or below 11km
#' (Mariana Trench).
#' \item GEODETIC_DATUM_ASSUMED_WGS84 Indicating that the interpreted coordinates assume they are
#' based on WGS84 datum as the datum was either not indicated or interpretable.
#' \item GEODETIC_DATUM_INVALID The geodetic datum given could not be interpreted.
#' \item IDENTIFIED_DATE_INVALID The date given for dwc:dateIdentified is invalid and cant be
#' interpreted at all.
#' \item IDENTIFIED_DATE_UNLIKELY The date given for dwc:dateIdentified is in the future or before
#' Linnean times (1700).
#' \item MODIFIED_DATE_INVALID A (partial) invalid date is given for dc:modified, such as a non
#' existing date, invalid zero month, etc.
#' \item MODIFIED_DATE_UNLIKELY The date given for dc:modified is in the future or predates unix
#' time (1970).
#' \item MULTIMEDIA_DATE_INVALID An invalid date is given for dc:created of a multimedia object.
#' \item MULTIMEDIA_URI_INVALID An invalid uri is given for a multimedia object.
#' \item PRESUMED_NEGATED_LATITUDE Latitude appears to be negated, e.g.
#' \item PRESUMED_NEGATED_LONGITUDE Longitude appears to be negated, e.g.
#' \item PRESUMED_SWAPPED_COORDINATE Latitude and longitude appear to be swapped.
#' \item RECORDED_DATE_INVALID A (partial) invalid date is given, such as a non existing date,
#' invalid zero month, etc.
#' \item RECORDED_DATE_MISMATCH The recording date specified as the eventDate string and the
#' individual year, month, day are contradicting.
#' \item RECORDED_DATE_UNLIKELY The recording date is highly unlikely, falling either into the
#' future or represents a very old date before 1600 that predates modern taxonomy.
#' \item REFERENCES_URI_INVALID An invalid uri is given for dc:references.
#' \item TAXON_MATCH_FUZZY Matching to the taxonomic backbone can only be done using a fuzzy, non
#' exact match.
#' \item TAXON_MATCH_HIGHERRANK Matching to the taxonomic backbone can only be done on a higher
#' rank and not the scientific name.
#' \item TAXON_MATCH_NONE Matching to the taxonomic backbone cannot be done cause there was no
#' match at all or several matches with too little information to keep them apart (homonyms).
#' \item TYPE_STATUS_INVALID The given type status is impossible to interpret or seriously
#' different from the recommended vocabulary.
#' \item ZERO_COORDINATE Coordinate is the exact 0/0 coordinate, often indicating a bad null
#' coordinate.
#' }
|
#' Analyse SAR TL measurements
#'
#' The function performs a SAR TL analysis on a
#' \code{\linkS4class{RLum.Analysis}} object including growth curve fitting.
#'
#' This function performs a SAR TL analysis on a set of curves. The SAR
#' procedure in general is given by Murray and Wintle (2000). For the
#' calculation of the Lx/Tx value the function \link{calc_TLLxTxRatio} is
#' used.\cr\cr \bold{Provided rejection criteria}\cr\cr
#' \sQuote{recyling.ratio}: calculated for every repeated regeneration dose
#' point.\cr \sQuote{recuperation.rate}: recuperation rate calculated by
#' comparing the Lx/Tx values of the zero regeneration point with the Ln/Tn
#' value (the Lx/Tx ratio of the natural signal). For methodological
#' background see Aitken and Smith (1988)\cr
#'
#' @param object \code{\linkS4class{RLum.Analysis}}(\bold{required}): input
#' object containing data for analysis
#'
#' @param object.background currently not used
#'
#' @param signal.integral.min \link{integer} (\bold{required}): requires the
#' channel number for the lower signal integral bound (e.g.
#' \code{signal.integral.min = 100})
#'
#' @param signal.integral.max \link{integer} (\bold{required}): requires the
#' channel number for the upper signal integral bound (e.g.
#' \code{signal.integral.max = 200})
#'
#' @param sequence.structure \link{vector} \link{character} (with default):
#' specifies the general sequence structure. Three steps are allowed (
#' \code{"PREHEAT"}, \code{"SIGNAL"}, \code{"BACKGROUND"}), in addition a
#' parameter \code{"EXCLUDE"}. This allows excluding TL curves which are not
#' relevant for the protocol analysis. (Note: None TL are removed by default)
#'
#' @param rejection.criteria \link{list} (with default): list containing
#' rejection criteria in percentage for the calculation.
#'
#' @param dose.points \code{\link{numeric}} (optional): option set dose points manually
#'
#' @param log \link{character} (with default): a character string which
#' contains "x" if the x axis is to be logarithmic, "y" if the y axis is to be
#' logarithmic and "xy" or "yx" if both axes are to be logarithmic. See
#' \link{plot.default}).
#'
#' @param \dots further arguments that will be passed to the function
#' \code{\link{plot_GrowthCurve}}
#'
#' @return A plot (optional) and an \code{\linkS4class{RLum.Results}} object is
#' returned containing the following elements:
#' \item{De.values}{\link{data.frame} containing De-values and further
#' parameters} \item{LnLxTnTx.values}{\link{data.frame} of all calculated Lx/Tx
#' values including signal, background counts and the dose points.}
#' \item{rejection.criteria}{\link{data.frame} with values that might by used
#' as rejection criteria. NA is produced if no R0 dose point exists.}\cr\cr
#' \bold{note:} the output should be accessed using the function
#' \code{\link{get_RLum}}
#' @note \bold{THIS IS A BETA VERSION}\cr\cr None TL curves will be removed
#' from the input object without further warning.
#' @section Function version: 0.1.4
#'
#' @author Sebastian Kreutzer, IRAMAT-CRP2A, Universite Bordeaux Montaigne (France)
#'
#' @seealso \code{\link{calc_TLLxTxRatio}}, \code{\link{plot_GrowthCurve}},
#' \code{\linkS4class{RLum.Analysis}}, \code{\linkS4class{RLum.Results}}
#' \code{\link{get_RLum}}
#'
#' @references Aitken, M.J. and Smith, B.W., 1988. Optical dating: recuperation
#' after bleaching. Quaternary Science Reviews 7, 387-393.
#'
#' Murray, A.S. and Wintle, A.G., 2000. Luminescence dating of quartz using an
#' improved single-aliquot regenerative-dose protocol. Radiation Measurements
#' 32, 57-73.
#' @keywords datagen plot
#' @examples
#'
#'
#' ##load data
#' data(ExampleData.BINfileData, envir = environment())
#'
#' ##transform the values from the first position in a RLum.Analysis object
#' object <- Risoe.BINfileData2RLum.Analysis(TL.SAR.Data, pos=3)
#'
#' ##perform analysis
#' analyse_SAR.TL(object,
#' signal.integral.min = 210,
#' signal.integral.max = 220,
#' log = "y",
#' fit.method = "EXP OR LIN",
#' sequence.structure = c("SIGNAL", "BACKGROUND"))
#'
#' @export
analyse_SAR.TL <- function(
object,
object.background,
signal.integral.min,
signal.integral.max,
sequence.structure = c("PREHEAT", "SIGNAL", "BACKGROUND"),
rejection.criteria = list(recycling.ratio = 10, recuperation.rate = 10),
dose.points,
log = "",
...
){
# CONFIG -----------------------------------------------------------------
##set allowed curve types
type.curves <- c("TL")
##=============================================================================#
# General Integrity Checks ---------------------------------------------------
##GENERAL
##MISSING INPUT
if(missing("object")==TRUE){
stop("[analyse_SAR.TL] No value set for 'object'!")
}
if(missing("signal.integral.min") == TRUE){
stop("[analyse_SAR.TL] No value set for 'signal.integral.min'!")
}
if(missing("signal.integral.max") == TRUE){
stop("[analyse_SAR.TL] No value set for 'signal.integral.max'!")
}
##INPUT OBJECTS
if(is(object, "RLum.Analysis") == FALSE){
stop("[analyse_SAR.TL] Input object is not of type 'RLum.Analyis'!")
}
# Protocol Integrity Checks --------------------------------------------------
##Remove non TL-curves from object by selecting TL curves
object@records <- get_RLum(object, recordType = type.curves)
##ANALYSE SEQUENCE OBJECT STRUCTURE
##set vector for sequence structure
temp.protocol.step <- rep(sequence.structure,length(object@records))[1:length(object@records)]
##grep object strucute
temp.sequence.structure <- structure_RLum(object)
##set values for step
temp.sequence.structure[,"protocol.step"] <- temp.protocol.step
##remove TL curves which are excluded
temp.sequence.structure <- temp.sequence.structure[which(
temp.sequence.structure[,"protocol.step"]!="EXCLUDE"),]
##check integrity; signal and bg range should be equal
if(length(
unique(
temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","x.max"]))>1){
stop(paste(
"[analyse_SAR.TL()] Signal range differs. Check sequence structure.\n",
temp.sequence.structure
))
}
##check if the wanted curves are a multiple of the structure
if(length(temp.sequence.structure[,"id"])%%length(sequence.structure)!=0){
stop("[analyse_SAR.TL()] Input TL curves are not a multiple of the sequence structure.")
}
# # Calculate LnLxTnTx values --------------------------------------------------
##grep IDs for signal and background curves
TL.preheat.ID <- temp.sequence.structure[
temp.sequence.structure[,"protocol.step"] == "PREHEAT","id"]
TL.signal.ID <- temp.sequence.structure[
temp.sequence.structure[,"protocol.step"] == "SIGNAL","id"]
TL.background.ID <- temp.sequence.structure[
temp.sequence.structure[,"protocol.step"] == "BACKGROUND","id"]
##calculate LxTx values using external function
for(i in seq(1,length(TL.signal.ID),by=2)){
temp.LnLxTnTx <- get_RLum(
calc_TLLxTxRatio(
Lx.data.signal = get_RLum(object, record.id=TL.signal.ID[i]),
Lx.data.background = get_RLum(object, record.id=TL.background.ID[i]),
Tx.data.signal = get_RLum(object, record.id=TL.signal.ID[i+1]),
Tx.data.background = get_RLum(object, record.id = TL.background.ID[i+1]),
signal.integral.min,
signal.integral.max))
##grep dose
temp.Dose <- object@records[[TL.signal.ID[i]]]@info$IRR_TIME
temp.LnLxTnTx <- cbind(Dose=temp.Dose, temp.LnLxTnTx)
if(exists("LnLxTnTx")==FALSE){
LnLxTnTx <- data.frame(temp.LnLxTnTx)
}else{
LnLxTnTx <- rbind(LnLxTnTx,temp.LnLxTnTx)
}
}
##set dose.points manual if argument was set
if(!missing(dose.points)){
temp.Dose <- dose.points
LnLxTnTx$Dose <- dose.points
}
# Set regeneration points -------------------------------------------------
#generate unique dose id - this are also the # for the generated points
temp.DoseID <- c(0:(length(temp.Dose)-1))
temp.DoseName <- paste("R",temp.DoseID,sep="")
temp.DoseName <- cbind(Name=temp.DoseName,Dose=temp.Dose)
##set natural
temp.DoseName[temp.DoseName[,"Name"]=="R0","Name"]<-"Natural"
##set R0
temp.DoseName[temp.DoseName[,"Name"]!="Natural" & temp.DoseName[,"Dose"]==0,"Name"]<-"R0"
##find duplicated doses (including 0 dose - which means the Natural)
temp.DoseDuplicated<-duplicated(temp.DoseName[,"Dose"])
##combine temp.DoseName
temp.DoseName<-cbind(temp.DoseName,Repeated=temp.DoseDuplicated)
##correct value for R0 (it is not really repeated)
temp.DoseName[temp.DoseName[,"Dose"]==0,"Repeated"]<-FALSE
##combine in the data frame
temp.LnLxTnTx<-data.frame(Name=temp.DoseName[,"Name"],
Repeated=as.logical(temp.DoseName[,"Repeated"]))
LnLxTnTx<-cbind(temp.LnLxTnTx,LnLxTnTx)
LnLxTnTx[,"Name"]<-as.character(LnLxTnTx[,"Name"])
# Calculate Recycling Ratio -----------------------------------------------
##Calculate Recycling Ratio
if(length(LnLxTnTx[LnLxTnTx[,"Repeated"]==TRUE,"Repeated"])>0){
##identify repeated doses
temp.Repeated<-LnLxTnTx[LnLxTnTx[,"Repeated"]==TRUE,c("Name","Dose","LxTx")]
##find concering previous dose for the repeated dose
temp.Previous<-t(sapply(1:length(temp.Repeated[,1]),function(x){
LnLxTnTx[LnLxTnTx[,"Dose"]==temp.Repeated[x,"Dose"] &
LnLxTnTx[,"Repeated"]==FALSE,c("Name","Dose","LxTx")]
}))
##convert to data.frame
temp.Previous<-as.data.frame(temp.Previous)
##set column names
temp.ColNames<-sapply(1:length(temp.Repeated[,1]),function(x){
paste(temp.Repeated[x,"Name"],"/",
temp.Previous[temp.Previous[,"Dose"]==temp.Repeated[x,"Dose"],"Name"],
sep="")
})
##Calculate Recycling Ratio
RecyclingRatio<-as.numeric(temp.Repeated[,"LxTx"])/as.numeric(temp.Previous[,"LxTx"])
##Just transform the matrix and add column names
RecyclingRatio<-t(RecyclingRatio)
colnames(RecyclingRatio)<-temp.ColNames
}else{RecyclingRatio<-NA}
# Calculate Recuperation Rate ---------------------------------------------
##Recuperation Rate
if("R0" %in% LnLxTnTx[,"Name"]==TRUE){
Recuperation<-round(LnLxTnTx[LnLxTnTx[,"Name"]=="R0","LxTx"]/
LnLxTnTx[LnLxTnTx[,"Name"]=="Natural","LxTx"],digits=4)
}else{Recuperation<-NA}
# Combine and Evaluate Rejection Criteria ---------------------------------
RejectionCriteria <- data.frame(
citeria = c(colnames(RecyclingRatio), "recuperation rate"),
value = c(RecyclingRatio,Recuperation),
threshold = c(
rep(paste("+/-", rejection.criteria$recycling.ratio/100)
,length(RecyclingRatio)),
paste("", rejection.criteria$recuperation.rate/100)
),
status = c(
if(is.na(RecyclingRatio)==FALSE){
sapply(1:length(RecyclingRatio), function(x){
if(abs(1-RecyclingRatio[x])>(rejection.criteria$recycling.ratio/100)){
"FAILED"
}else{"OK"}})}else{NA},
if(is.na(Recuperation)==FALSE &
Recuperation>rejection.criteria$recuperation.rate){"FAILED"}else{"OK"}
))
##============================================================================##
##PLOTTING
##============================================================================##
# Plotting - Config -------------------------------------------------------
##grep plot parameter
par.default <- par(no.readonly = TRUE)
##colours and double for plotting
col <- get("col", pos = .LuminescenceEnv)
col.doubled <- rep(col, each=2)
layout(matrix(c(1,1,2,2,
1,1,2,2,
3,3,4,4,
3,3,4,4,
5,5,5,5),5,4,byrow=TRUE))
par(oma=c(0,0,0,0), mar=c(4,4,3,3))
## 1 -> TL Lx
## 2 -> TL Tx
## 3 -> TL Lx Plateau
## 4 -> TL Tx Plateau
## 5 -> Legend
##recalculate signal.integral from channels to temperature
signal.integral.temperature <- c(object@records[[TL.signal.ID[1]]]@data[signal.integral.min,1] :
object@records[[TL.signal.ID[1]]]@data[signal.integral.max,1])
##warning if number of curves exceed colour values
if(length(col)<length(TL.signal.ID/2)){
cat("\n[analyse_SAR.TL.R] Warning: To many curves! Only the first",
length(col),"curves are plotted!")
}
# # Plotting TL Lx Curves ----------------------------------------------------
#open plot area LnLx
plot(NA,NA,
xlab="Temp. [\u00B0C]",
ylab=paste("TL [a.u.]",sep=""),
xlim=c(0.1,
max(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","x.max"])),
ylim=c(
min(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","y.min"]),
max(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","y.max"])),
main=expression(paste(L[n],",",L[x]," curves",sep="")),
log=log)
##plot curves
sapply(seq(1,length(TL.signal.ID),by=2), function(x){
lines(object@records[[TL.signal.ID[x]]]@data,col=col.doubled[x])
})
##mark integration limits
abline(v=min(signal.integral.temperature), lty=2, col="gray")
abline(v=max(signal.integral.temperature), lty=2, col="gray")
# Plotting TnTx Curves ----------------------------------------------------
#open plot area TnTx
plot(NA,NA,
xlab="Temp. [\u00B0C]",
ylab=paste("TL [a.u.]",sep=""),
xlim=c(0.1,
max(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","x.max"])),
ylim=c(
min(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","y.min"]),
max(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","y.max"])),
main=expression(paste(T[n],",",T[x]," curves",sep="")),
log=log)
##plot curves
sapply(seq(2,length(TL.signal.ID),by=2), function(x){
lines(object@records[[TL.signal.ID[x]]]@data,col=col.doubled[x])
})
##mark integration limits
abline(v=min(signal.integral.temperature), lty=2, col="gray")
abline(v=max(signal.integral.temperature), lty=2, col="gray")
# Plotting Plateau Test LnLx -------------------------------------------------
NTL.net.LnLx <- data.frame(object@records[[TL.signal.ID[1]]]@data[,1],
object@records[[TL.signal.ID[1]]]@data[,2]-
object@records[[TL.background.ID[1]]]@data[,2])
Reg1.net.LnLx <- data.frame(object@records[[TL.signal.ID[3]]]@data[,1],
object@records[[TL.signal.ID[3]]]@data[,2]-
object@records[[TL.background.ID[3]]]@data[,2])
TL.Plateau.LnLx <- data.frame(NTL.net.LnLx[,1], Reg1.net.LnLx[,2]/NTL.net.LnLx[,2])
##Plot Plateau Test
plot(NA, NA,
xlab = "Temp. [\u00B0C]",
ylab = "TL [a.u.]",
xlim = c(min(signal.integral.temperature)*0.9, max(signal.integral.temperature)*1.1),
ylim = c(0, max(NTL.net.LnLx[,2])),
main = expression(paste("Plateau test ",L[n],",",L[x]," curves",sep=""))
)
##plot single curves
lines(NTL.net.LnLx, col=col[1])
lines(Reg1.net.LnLx, col=col[2])
##plot
par(new=TRUE)
plot(TL.Plateau.LnLx,
axes=FALSE,
xlab="",
ylab="",
ylim=c(0,
quantile(TL.Plateau.LnLx[c(signal.integral.min:signal.integral.max),2],
probs = c(0.90), na.rm = TRUE)+3),
col="darkgreen")
axis(4)
# Plotting Plateau Test TnTx -------------------------------------------------
##get NTL signal
NTL.net.TnTx <- data.frame(object@records[[TL.signal.ID[2]]]@data[,1],
object@records[[TL.signal.ID[2]]]@data[,2]-
object@records[[TL.background.ID[2]]]@data[,2])
##get signal from the first regeneration point
Reg1.net.TnTx <- data.frame(object@records[[TL.signal.ID[4]]]@data[,1],
object@records[[TL.signal.ID[4]]]@data[,2]-
object@records[[TL.background.ID[4]]]@data[,2])
##combine values
TL.Plateau.TnTx <- data.frame(NTL.net.TnTx[,1], Reg1.net.TnTx[,2]/NTL.net.TnTx[,2])
##Plot Plateau Test
plot(NA, NA,
xlab = "Temp. [\u00B0C]",
ylab = "TL [a.u.]",
xlim = c(min(signal.integral.temperature)*0.9, max(signal.integral.temperature)*1.1),
ylim = c(0, max(NTL.net.TnTx[,2])),
main = expression(paste("plateau Test ",T[n],",",T[x]," curves",sep=""))
)
##plot single curves
lines(NTL.net.TnTx, col=col[1])
lines(Reg1.net.TnTx, col=col[2])
##plot
par(new=TRUE)
plot(TL.Plateau.TnTx,
axes=FALSE,
xlab="",
ylab="",
ylim=c(0,
quantile(TL.Plateau.TnTx[c(signal.integral.min:signal.integral.max),2],
probs = c(0.90), na.rm = TRUE)+3),
col="darkgreen")
axis(4)
# Plotting Legend ----------------------------------------
plot(c(1:(length(TL.signal.ID)/2)),
rep(8,length(TL.signal.ID)/2),
type = "p",
axes=FALSE,
xlab="",
ylab="",
pch=15,
col=col[1:length(TL.signal.ID)],
cex=2,
ylim=c(0,10)
)
##add text
text(c(1:(length(TL.signal.ID)/2)),
rep(4,length(TL.signal.ID)/2),
paste(LnLxTnTx$Name,"\n(",LnLxTnTx$Dose,")", sep="")
)
##add line
abline(h=10,lwd=0.5)
##set failed text and mark De as failed
if(length(grep("FAILED",RejectionCriteria$status))>0){
mtext("[FAILED]", col="red")
}
##reset par
par(par.default)
rm(par.default)
# Plotting GC ----------------------------------------
temp.sample <- data.frame(Dose=LnLxTnTx$Dose,
LxTx=LnLxTnTx$LxTx,
LxTx.Error=LnLxTnTx$LxTx*0.1,
TnTx=LnLxTnTx$TnTx
)
temp.GC <- get_RLum(plot_GrowthCurve(temp.sample,
...))[,c("De","De.Error")]
##add recjection status
if(length(grep("FAILED",RejectionCriteria$status))>0){
temp.GC <- data.frame(temp.GC, RC.Status="FAILED")
}else{
temp.GC <- data.frame(temp.GC, RC.Status="OK")
}
# Return Values -----------------------------------------------------------
newRLumResults.analyse_SAR.TL <- set_RLum(
class = "RLum.Results",
data = list(
De.values = temp.GC,
LnLxTnTx.table = LnLxTnTx,
rejection.criteria = RejectionCriteria))
return(newRLumResults.analyse_SAR.TL)
}
| /Luminescence/R/analyse_SAR.TL.R | no_license | ingted/R-Examples | R | false | false | 18,691 | r | #' Analyse SAR TL measurements
#'
#' The function performs a SAR TL analysis on a
#' \code{\linkS4class{RLum.Analysis}} object including growth curve fitting.
#'
#' This function performs a SAR TL analysis on a set of curves. The SAR
#' procedure in general is given by Murray and Wintle (2000). For the
#' calculation of the Lx/Tx value the function \link{calc_TLLxTxRatio} is
#' used.\cr\cr \bold{Provided rejection criteria}\cr\cr
#' \sQuote{recyling.ratio}: calculated for every repeated regeneration dose
#' point.\cr \sQuote{recuperation.rate}: recuperation rate calculated by
#' comparing the Lx/Tx values of the zero regeneration point with the Ln/Tn
#' value (the Lx/Tx ratio of the natural signal). For methodological
#' background see Aitken and Smith (1988)\cr
#'
#' @param object \code{\linkS4class{RLum.Analysis}}(\bold{required}): input
#' object containing data for analysis
#'
#' @param object.background currently not used
#'
#' @param signal.integral.min \link{integer} (\bold{required}): requires the
#' channel number for the lower signal integral bound (e.g.
#' \code{signal.integral.min = 100})
#'
#' @param signal.integral.max \link{integer} (\bold{required}): requires the
#' channel number for the upper signal integral bound (e.g.
#' \code{signal.integral.max = 200})
#'
#' @param sequence.structure \link{vector} \link{character} (with default):
#' specifies the general sequence structure. Three steps are allowed (
#' \code{"PREHEAT"}, \code{"SIGNAL"}, \code{"BACKGROUND"}), in addition a
#' parameter \code{"EXCLUDE"}. This allows excluding TL curves which are not
#' relevant for the protocol analysis. (Note: None TL are removed by default)
#'
#' @param rejection.criteria \link{list} (with default): list containing
#' rejection criteria in percentage for the calculation.
#'
#' @param dose.points \code{\link{numeric}} (optional): option set dose points manually
#'
#' @param log \link{character} (with default): a character string which
#' contains "x" if the x axis is to be logarithmic, "y" if the y axis is to be
#' logarithmic and "xy" or "yx" if both axes are to be logarithmic. See
#' \link{plot.default}).
#'
#' @param \dots further arguments that will be passed to the function
#' \code{\link{plot_GrowthCurve}}
#'
#' @return A plot (optional) and an \code{\linkS4class{RLum.Results}} object is
#' returned containing the following elements:
#' \item{De.values}{\link{data.frame} containing De-values and further
#' parameters} \item{LnLxTnTx.values}{\link{data.frame} of all calculated Lx/Tx
#' values including signal, background counts and the dose points.}
#' \item{rejection.criteria}{\link{data.frame} with values that might by used
#' as rejection criteria. NA is produced if no R0 dose point exists.}\cr\cr
#' \bold{note:} the output should be accessed using the function
#' \code{\link{get_RLum}}
#' @note \bold{THIS IS A BETA VERSION}\cr\cr None TL curves will be removed
#' from the input object without further warning.
#' @section Function version: 0.1.4
#'
#' @author Sebastian Kreutzer, IRAMAT-CRP2A, Universite Bordeaux Montaigne (France)
#'
#' @seealso \code{\link{calc_TLLxTxRatio}}, \code{\link{plot_GrowthCurve}},
#' \code{\linkS4class{RLum.Analysis}}, \code{\linkS4class{RLum.Results}}
#' \code{\link{get_RLum}}
#'
#' @references Aitken, M.J. and Smith, B.W., 1988. Optical dating: recuperation
#' after bleaching. Quaternary Science Reviews 7, 387-393.
#'
#' Murray, A.S. and Wintle, A.G., 2000. Luminescence dating of quartz using an
#' improved single-aliquot regenerative-dose protocol. Radiation Measurements
#' 32, 57-73.
#' @keywords datagen plot
#' @examples
#'
#'
#' ##load data
#' data(ExampleData.BINfileData, envir = environment())
#'
#' ##transform the values from the first position in a RLum.Analysis object
#' object <- Risoe.BINfileData2RLum.Analysis(TL.SAR.Data, pos=3)
#'
#' ##perform analysis
#' analyse_SAR.TL(object,
#' signal.integral.min = 210,
#' signal.integral.max = 220,
#' log = "y",
#' fit.method = "EXP OR LIN",
#' sequence.structure = c("SIGNAL", "BACKGROUND"))
#'
#' @export
analyse_SAR.TL <- function(
object,
object.background,
signal.integral.min,
signal.integral.max,
sequence.structure = c("PREHEAT", "SIGNAL", "BACKGROUND"),
rejection.criteria = list(recycling.ratio = 10, recuperation.rate = 10),
dose.points,
log = "",
...
){
# CONFIG -----------------------------------------------------------------
##set allowed curve types
type.curves <- c("TL")
##=============================================================================#
# General Integrity Checks ---------------------------------------------------
##GENERAL
##MISSING INPUT
if(missing("object")==TRUE){
stop("[analyse_SAR.TL] No value set for 'object'!")
}
if(missing("signal.integral.min") == TRUE){
stop("[analyse_SAR.TL] No value set for 'signal.integral.min'!")
}
if(missing("signal.integral.max") == TRUE){
stop("[analyse_SAR.TL] No value set for 'signal.integral.max'!")
}
##INPUT OBJECTS
if(is(object, "RLum.Analysis") == FALSE){
stop("[analyse_SAR.TL] Input object is not of type 'RLum.Analyis'!")
}
# Protocol Integrity Checks --------------------------------------------------
##Remove non TL-curves from object by selecting TL curves
object@records <- get_RLum(object, recordType = type.curves)
##ANALYSE SEQUENCE OBJECT STRUCTURE
##set vector for sequence structure
temp.protocol.step <- rep(sequence.structure,length(object@records))[1:length(object@records)]
##grep object strucute
temp.sequence.structure <- structure_RLum(object)
##set values for step
temp.sequence.structure[,"protocol.step"] <- temp.protocol.step
##remove TL curves which are excluded
temp.sequence.structure <- temp.sequence.structure[which(
temp.sequence.structure[,"protocol.step"]!="EXCLUDE"),]
##check integrity; signal and bg range should be equal
if(length(
unique(
temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","x.max"]))>1){
stop(paste(
"[analyse_SAR.TL()] Signal range differs. Check sequence structure.\n",
temp.sequence.structure
))
}
##check if the wanted curves are a multiple of the structure
if(length(temp.sequence.structure[,"id"])%%length(sequence.structure)!=0){
stop("[analyse_SAR.TL()] Input TL curves are not a multiple of the sequence structure.")
}
# # Calculate LnLxTnTx values --------------------------------------------------
##grep IDs for signal and background curves
TL.preheat.ID <- temp.sequence.structure[
temp.sequence.structure[,"protocol.step"] == "PREHEAT","id"]
TL.signal.ID <- temp.sequence.structure[
temp.sequence.structure[,"protocol.step"] == "SIGNAL","id"]
TL.background.ID <- temp.sequence.structure[
temp.sequence.structure[,"protocol.step"] == "BACKGROUND","id"]
##calculate LxTx values using external function
for(i in seq(1,length(TL.signal.ID),by=2)){
temp.LnLxTnTx <- get_RLum(
calc_TLLxTxRatio(
Lx.data.signal = get_RLum(object, record.id=TL.signal.ID[i]),
Lx.data.background = get_RLum(object, record.id=TL.background.ID[i]),
Tx.data.signal = get_RLum(object, record.id=TL.signal.ID[i+1]),
Tx.data.background = get_RLum(object, record.id = TL.background.ID[i+1]),
signal.integral.min,
signal.integral.max))
##grep dose
temp.Dose <- object@records[[TL.signal.ID[i]]]@info$IRR_TIME
temp.LnLxTnTx <- cbind(Dose=temp.Dose, temp.LnLxTnTx)
if(exists("LnLxTnTx")==FALSE){
LnLxTnTx <- data.frame(temp.LnLxTnTx)
}else{
LnLxTnTx <- rbind(LnLxTnTx,temp.LnLxTnTx)
}
}
##set dose.points manual if argument was set
if(!missing(dose.points)){
temp.Dose <- dose.points
LnLxTnTx$Dose <- dose.points
}
# Set regeneration points -------------------------------------------------
#generate unique dose id - this are also the # for the generated points
temp.DoseID <- c(0:(length(temp.Dose)-1))
temp.DoseName <- paste("R",temp.DoseID,sep="")
temp.DoseName <- cbind(Name=temp.DoseName,Dose=temp.Dose)
##set natural
temp.DoseName[temp.DoseName[,"Name"]=="R0","Name"]<-"Natural"
##set R0
temp.DoseName[temp.DoseName[,"Name"]!="Natural" & temp.DoseName[,"Dose"]==0,"Name"]<-"R0"
##find duplicated doses (including 0 dose - which means the Natural)
temp.DoseDuplicated<-duplicated(temp.DoseName[,"Dose"])
##combine temp.DoseName
temp.DoseName<-cbind(temp.DoseName,Repeated=temp.DoseDuplicated)
##correct value for R0 (it is not really repeated)
temp.DoseName[temp.DoseName[,"Dose"]==0,"Repeated"]<-FALSE
##combine in the data frame
temp.LnLxTnTx<-data.frame(Name=temp.DoseName[,"Name"],
Repeated=as.logical(temp.DoseName[,"Repeated"]))
LnLxTnTx<-cbind(temp.LnLxTnTx,LnLxTnTx)
LnLxTnTx[,"Name"]<-as.character(LnLxTnTx[,"Name"])
# Calculate Recycling Ratio -----------------------------------------------
##Calculate Recycling Ratio
if(length(LnLxTnTx[LnLxTnTx[,"Repeated"]==TRUE,"Repeated"])>0){
##identify repeated doses
temp.Repeated<-LnLxTnTx[LnLxTnTx[,"Repeated"]==TRUE,c("Name","Dose","LxTx")]
##find concering previous dose for the repeated dose
temp.Previous<-t(sapply(1:length(temp.Repeated[,1]),function(x){
LnLxTnTx[LnLxTnTx[,"Dose"]==temp.Repeated[x,"Dose"] &
LnLxTnTx[,"Repeated"]==FALSE,c("Name","Dose","LxTx")]
}))
##convert to data.frame
temp.Previous<-as.data.frame(temp.Previous)
##set column names
temp.ColNames<-sapply(1:length(temp.Repeated[,1]),function(x){
paste(temp.Repeated[x,"Name"],"/",
temp.Previous[temp.Previous[,"Dose"]==temp.Repeated[x,"Dose"],"Name"],
sep="")
})
##Calculate Recycling Ratio
RecyclingRatio<-as.numeric(temp.Repeated[,"LxTx"])/as.numeric(temp.Previous[,"LxTx"])
##Just transform the matrix and add column names
RecyclingRatio<-t(RecyclingRatio)
colnames(RecyclingRatio)<-temp.ColNames
}else{RecyclingRatio<-NA}
# Calculate Recuperation Rate ---------------------------------------------
##Recuperation Rate
if("R0" %in% LnLxTnTx[,"Name"]==TRUE){
Recuperation<-round(LnLxTnTx[LnLxTnTx[,"Name"]=="R0","LxTx"]/
LnLxTnTx[LnLxTnTx[,"Name"]=="Natural","LxTx"],digits=4)
}else{Recuperation<-NA}
# Combine and Evaluate Rejection Criteria ---------------------------------
RejectionCriteria <- data.frame(
citeria = c(colnames(RecyclingRatio), "recuperation rate"),
value = c(RecyclingRatio,Recuperation),
threshold = c(
rep(paste("+/-", rejection.criteria$recycling.ratio/100)
,length(RecyclingRatio)),
paste("", rejection.criteria$recuperation.rate/100)
),
status = c(
if(is.na(RecyclingRatio)==FALSE){
sapply(1:length(RecyclingRatio), function(x){
if(abs(1-RecyclingRatio[x])>(rejection.criteria$recycling.ratio/100)){
"FAILED"
}else{"OK"}})}else{NA},
if(is.na(Recuperation)==FALSE &
Recuperation>rejection.criteria$recuperation.rate){"FAILED"}else{"OK"}
))
##============================================================================##
##PLOTTING
##============================================================================##
# Plotting - Config -------------------------------------------------------
##grep plot parameter
par.default <- par(no.readonly = TRUE)
##colours and double for plotting
col <- get("col", pos = .LuminescenceEnv)
col.doubled <- rep(col, each=2)
layout(matrix(c(1,1,2,2,
1,1,2,2,
3,3,4,4,
3,3,4,4,
5,5,5,5),5,4,byrow=TRUE))
par(oma=c(0,0,0,0), mar=c(4,4,3,3))
## 1 -> TL Lx
## 2 -> TL Tx
## 3 -> TL Lx Plateau
## 4 -> TL Tx Plateau
## 5 -> Legend
##recalculate signal.integral from channels to temperature
signal.integral.temperature <- c(object@records[[TL.signal.ID[1]]]@data[signal.integral.min,1] :
object@records[[TL.signal.ID[1]]]@data[signal.integral.max,1])
##warning if number of curves exceed colour values
if(length(col)<length(TL.signal.ID/2)){
cat("\n[analyse_SAR.TL.R] Warning: To many curves! Only the first",
length(col),"curves are plotted!")
}
# # Plotting TL Lx Curves ----------------------------------------------------
#open plot area LnLx
plot(NA,NA,
xlab="Temp. [\u00B0C]",
ylab=paste("TL [a.u.]",sep=""),
xlim=c(0.1,
max(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","x.max"])),
ylim=c(
min(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","y.min"]),
max(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","y.max"])),
main=expression(paste(L[n],",",L[x]," curves",sep="")),
log=log)
##plot curves
sapply(seq(1,length(TL.signal.ID),by=2), function(x){
lines(object@records[[TL.signal.ID[x]]]@data,col=col.doubled[x])
})
##mark integration limits
abline(v=min(signal.integral.temperature), lty=2, col="gray")
abline(v=max(signal.integral.temperature), lty=2, col="gray")
# Plotting TnTx Curves ----------------------------------------------------
#open plot area TnTx
plot(NA,NA,
xlab="Temp. [\u00B0C]",
ylab=paste("TL [a.u.]",sep=""),
xlim=c(0.1,
max(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","x.max"])),
ylim=c(
min(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","y.min"]),
max(temp.sequence.structure[temp.sequence.structure[,"protocol.step"]=="SIGNAL","y.max"])),
main=expression(paste(T[n],",",T[x]," curves",sep="")),
log=log)
##plot curves
sapply(seq(2,length(TL.signal.ID),by=2), function(x){
lines(object@records[[TL.signal.ID[x]]]@data,col=col.doubled[x])
})
##mark integration limits
abline(v=min(signal.integral.temperature), lty=2, col="gray")
abline(v=max(signal.integral.temperature), lty=2, col="gray")
# Plotting Plateau Test LnLx -------------------------------------------------
NTL.net.LnLx <- data.frame(object@records[[TL.signal.ID[1]]]@data[,1],
object@records[[TL.signal.ID[1]]]@data[,2]-
object@records[[TL.background.ID[1]]]@data[,2])
Reg1.net.LnLx <- data.frame(object@records[[TL.signal.ID[3]]]@data[,1],
object@records[[TL.signal.ID[3]]]@data[,2]-
object@records[[TL.background.ID[3]]]@data[,2])
TL.Plateau.LnLx <- data.frame(NTL.net.LnLx[,1], Reg1.net.LnLx[,2]/NTL.net.LnLx[,2])
##Plot Plateau Test
plot(NA, NA,
xlab = "Temp. [\u00B0C]",
ylab = "TL [a.u.]",
xlim = c(min(signal.integral.temperature)*0.9, max(signal.integral.temperature)*1.1),
ylim = c(0, max(NTL.net.LnLx[,2])),
main = expression(paste("Plateau test ",L[n],",",L[x]," curves",sep=""))
)
##plot single curves
lines(NTL.net.LnLx, col=col[1])
lines(Reg1.net.LnLx, col=col[2])
##plot
par(new=TRUE)
plot(TL.Plateau.LnLx,
axes=FALSE,
xlab="",
ylab="",
ylim=c(0,
quantile(TL.Plateau.LnLx[c(signal.integral.min:signal.integral.max),2],
probs = c(0.90), na.rm = TRUE)+3),
col="darkgreen")
axis(4)
# Plotting Plateau Test TnTx -------------------------------------------------
##get NTL signal
NTL.net.TnTx <- data.frame(object@records[[TL.signal.ID[2]]]@data[,1],
object@records[[TL.signal.ID[2]]]@data[,2]-
object@records[[TL.background.ID[2]]]@data[,2])
##get signal from the first regeneration point
Reg1.net.TnTx <- data.frame(object@records[[TL.signal.ID[4]]]@data[,1],
object@records[[TL.signal.ID[4]]]@data[,2]-
object@records[[TL.background.ID[4]]]@data[,2])
##combine values
TL.Plateau.TnTx <- data.frame(NTL.net.TnTx[,1], Reg1.net.TnTx[,2]/NTL.net.TnTx[,2])
##Plot Plateau Test
plot(NA, NA,
xlab = "Temp. [\u00B0C]",
ylab = "TL [a.u.]",
xlim = c(min(signal.integral.temperature)*0.9, max(signal.integral.temperature)*1.1),
ylim = c(0, max(NTL.net.TnTx[,2])),
main = expression(paste("plateau Test ",T[n],",",T[x]," curves",sep=""))
)
##plot single curves
lines(NTL.net.TnTx, col=col[1])
lines(Reg1.net.TnTx, col=col[2])
##plot
par(new=TRUE)
plot(TL.Plateau.TnTx,
axes=FALSE,
xlab="",
ylab="",
ylim=c(0,
quantile(TL.Plateau.TnTx[c(signal.integral.min:signal.integral.max),2],
probs = c(0.90), na.rm = TRUE)+3),
col="darkgreen")
axis(4)
# Plotting Legend ----------------------------------------
plot(c(1:(length(TL.signal.ID)/2)),
rep(8,length(TL.signal.ID)/2),
type = "p",
axes=FALSE,
xlab="",
ylab="",
pch=15,
col=col[1:length(TL.signal.ID)],
cex=2,
ylim=c(0,10)
)
##add text
text(c(1:(length(TL.signal.ID)/2)),
rep(4,length(TL.signal.ID)/2),
paste(LnLxTnTx$Name,"\n(",LnLxTnTx$Dose,")", sep="")
)
##add line
abline(h=10,lwd=0.5)
##set failed text and mark De as failed
if(length(grep("FAILED",RejectionCriteria$status))>0){
mtext("[FAILED]", col="red")
}
##reset par
par(par.default)
rm(par.default)
# Plotting GC ----------------------------------------
temp.sample <- data.frame(Dose=LnLxTnTx$Dose,
LxTx=LnLxTnTx$LxTx,
LxTx.Error=LnLxTnTx$LxTx*0.1,
TnTx=LnLxTnTx$TnTx
)
temp.GC <- get_RLum(plot_GrowthCurve(temp.sample,
...))[,c("De","De.Error")]
##add recjection status
if(length(grep("FAILED",RejectionCriteria$status))>0){
temp.GC <- data.frame(temp.GC, RC.Status="FAILED")
}else{
temp.GC <- data.frame(temp.GC, RC.Status="OK")
}
# Return Values -----------------------------------------------------------
newRLumResults.analyse_SAR.TL <- set_RLum(
class = "RLum.Results",
data = list(
De.values = temp.GC,
LnLxTnTx.table = LnLxTnTx,
rejection.criteria = RejectionCriteria))
return(newRLumResults.analyse_SAR.TL)
}
|
complete <- function(directory, id = 1:332){
data <- NULL
result <- NULL
# get a list of all files in the specified directory that end in .csv
files <- list.files(directory, pattern="*.csv")
# loop through the monitors specified
for(index in id)
{
# build a path to the file
path <- file.path(directory, files[index])
# print(path)
# read in the data
data <- read.csv(path)#, nrows=5)
# only include complete cases
data <- data[complete.cases(data),]
# combine, by rows, the monitor number and a sum of the completed cases
result <- rbind(result, c(id=index, nobs=nrow(data)))
}
# print(result)
# print(class(result))
# requirements state to return this result as a dataframe (convert our matrix to a dataframe)
return(data.frame(result))
}
answer <- complete("/home/user/Downloads/specdata", 1)
print(answer)
print("=====")
answer <- complete("/home/user/Downloads/specdata", c(2, 4, 8, 10, 12))
print(answer)
print("=====")
answer <- complete("/home/user/Downloads/specdata", 30:25)
print(answer)
print("=====")
answer <- complete("/home/user/Downloads/specdata", 3)
print(answer)
| /Week2/complete.R | no_license | XAGV1YBGAdk34WDPVVLn/datasciencecoursera | R | false | false | 1,212 | r | complete <- function(directory, id = 1:332){
data <- NULL
result <- NULL
# get a list of all files in the specified directory that end in .csv
files <- list.files(directory, pattern="*.csv")
# loop through the monitors specified
for(index in id)
{
# build a path to the file
path <- file.path(directory, files[index])
# print(path)
# read in the data
data <- read.csv(path)#, nrows=5)
# only include complete cases
data <- data[complete.cases(data),]
# combine, by rows, the monitor number and a sum of the completed cases
result <- rbind(result, c(id=index, nobs=nrow(data)))
}
# print(result)
# print(class(result))
# requirements state to return this result as a dataframe (convert our matrix to a dataframe)
return(data.frame(result))
}
answer <- complete("/home/user/Downloads/specdata", 1)
print(answer)
print("=====")
answer <- complete("/home/user/Downloads/specdata", c(2, 4, 8, 10, 12))
print(answer)
print("=====")
answer <- complete("/home/user/Downloads/specdata", 30:25)
print(answer)
print("=====")
answer <- complete("/home/user/Downloads/specdata", 3)
print(answer)
|
# Set working direcotory
this.dir <- dirname(parent.frame(2)$ofile)
setwd(this.dir)
library(dplyr)
library(rpart)
library(rpart.plot)
library(party)
library(readr)
#Feature Engineering
train <- read_csv('HR_comma_sep.csv')
train$sales <- as.factor(train$sales)
train$salary <- as.factor(train$salary)
#Fitting a Conditional Inference Tree
set.seed <- (1987)
fit <- cforest(as.factor(left) ~ .,
data = train,
controls = cforest_unbiased(ntree = 2000, mtry = 3))
#Viewing a sample tree
party:::prettytree(fit@ensemble[[1]], names(fit@data@get("input")))
#Predicting for the test dataset using CIT
prediction <- predict(fit, train, OOB = TRUE, type = "response")
submit <- data.frame(Actual = train$left,
Predicted = prediction)
prop.table(table(submit$Actual, submit$Predicted),1)
| /HR Analytics/HR_Analytics.R | no_license | kshirasaagar/Kaggle | R | false | false | 841 | r |
# Set working direcotory
this.dir <- dirname(parent.frame(2)$ofile)
setwd(this.dir)
library(dplyr)
library(rpart)
library(rpart.plot)
library(party)
library(readr)
#Feature Engineering
train <- read_csv('HR_comma_sep.csv')
train$sales <- as.factor(train$sales)
train$salary <- as.factor(train$salary)
#Fitting a Conditional Inference Tree
set.seed <- (1987)
fit <- cforest(as.factor(left) ~ .,
data = train,
controls = cforest_unbiased(ntree = 2000, mtry = 3))
#Viewing a sample tree
party:::prettytree(fit@ensemble[[1]], names(fit@data@get("input")))
#Predicting for the test dataset using CIT
prediction <- predict(fit, train, OOB = TRUE, type = "response")
submit <- data.frame(Actual = train$left,
Predicted = prediction)
prop.table(table(submit$Actual, submit$Predicted),1)
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/lime-package.r
\docType{package}
\name{lime-package}
\alias{lime-package}
\alias{_PACKAGE}
\title{lime: Local Interpretable Model-Agnostic Explanations}
\description{
\if{html}{\figure{logo.png}{options: align='right'}}
When building complex models, it is often difficult to explain why
the model should be trusted. While global measures such as accuracy are
useful, they cannot be used for explaining why a model made a specific
prediction. 'lime' (a port of the 'lime' 'Python' package) is a method for
explaining the outcome of black box models by fitting a local model around
the point in question an perturbations of this point. The approach is
described in more detail in the article by Ribeiro et al. (2016)
<arXiv:1602.04938>.
}
\details{
This package is a port of the original Python lime package implementing the
prediction explanation framework laid out Ribeiro \emph{et al.} (2016). The
package supports models from \code{caret} and \code{mlr} natively, but see
\link[=model_support]{the docs} for how to make it work for any model.
\strong{Main functions:}
Use of \code{lime} is mainly through two functions. First you create an
\code{explainer} object using the \code{\link[=lime]{lime()}} function based on the training data and
the model, and then you can use the \code{\link[=explain]{explain()}} function along with new data
and the explainer to create explanations for the model output.
Along with these two functions, \code{lime} also provides the \code{\link[=plot_features]{plot_features()}}
and \code{\link[=plot_text_explanations]{plot_text_explanations()}} function to visualise the explanations
directly.
}
\references{
Ribeiro, M.T., Singh, S., Guestrin, C. \emph{"Why Should I Trust You?": Explaining the Predictions of Any Classifier}. 2016, \url{https://arxiv.org/abs/1602.04938}
}
\seealso{
Useful links:
\itemize{
\item \url{https://lime.data-imaginist.com}
\item \url{https://github.com/thomasp85/lime}
\item Report bugs at \url{https://github.com/thomasp85/lime/issues}
}
}
\author{
\strong{Maintainer}: Thomas Lin Pedersen \email{thomasp85@gmail.com} (0000-0002-5147-4711)
Authors:
\itemize{
\item MichaΓ«l Benesty \email{michael@benesty.fr}
}
}
| /man/lime-package.Rd | permissive | goodekat/lime | R | false | true | 2,303 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/lime-package.r
\docType{package}
\name{lime-package}
\alias{lime-package}
\alias{_PACKAGE}
\title{lime: Local Interpretable Model-Agnostic Explanations}
\description{
\if{html}{\figure{logo.png}{options: align='right'}}
When building complex models, it is often difficult to explain why
the model should be trusted. While global measures such as accuracy are
useful, they cannot be used for explaining why a model made a specific
prediction. 'lime' (a port of the 'lime' 'Python' package) is a method for
explaining the outcome of black box models by fitting a local model around
the point in question an perturbations of this point. The approach is
described in more detail in the article by Ribeiro et al. (2016)
<arXiv:1602.04938>.
}
\details{
This package is a port of the original Python lime package implementing the
prediction explanation framework laid out Ribeiro \emph{et al.} (2016). The
package supports models from \code{caret} and \code{mlr} natively, but see
\link[=model_support]{the docs} for how to make it work for any model.
\strong{Main functions:}
Use of \code{lime} is mainly through two functions. First you create an
\code{explainer} object using the \code{\link[=lime]{lime()}} function based on the training data and
the model, and then you can use the \code{\link[=explain]{explain()}} function along with new data
and the explainer to create explanations for the model output.
Along with these two functions, \code{lime} also provides the \code{\link[=plot_features]{plot_features()}}
and \code{\link[=plot_text_explanations]{plot_text_explanations()}} function to visualise the explanations
directly.
}
\references{
Ribeiro, M.T., Singh, S., Guestrin, C. \emph{"Why Should I Trust You?": Explaining the Predictions of Any Classifier}. 2016, \url{https://arxiv.org/abs/1602.04938}
}
\seealso{
Useful links:
\itemize{
\item \url{https://lime.data-imaginist.com}
\item \url{https://github.com/thomasp85/lime}
\item Report bugs at \url{https://github.com/thomasp85/lime/issues}
}
}
\author{
\strong{Maintainer}: Thomas Lin Pedersen \email{thomasp85@gmail.com} (0000-0002-5147-4711)
Authors:
\itemize{
\item MichaΓ«l Benesty \email{michael@benesty.fr}
}
}
|
\name{calendarHeat}
\alias{calendarHeat}
\title{An R function to display time-series data as a calendar heatmap}
\usage{
calendarHeat(dates, values, ncolors = 99, color = "r2g",
varname = "Values", date.form = "\%Y-\%m-\%d", ...)
}
\arguments{
\item{dates}{vector of Dates}
\item{values}{vector of values}
\item{ncolors}{number of colors to use}
\item{color}{the color scheme to use. Currently supports
r2b, (red to blue), r2g (red to green), and w2b (white to
blue).}
\item{varname}{varaible names}
\item{date.form}{the format of the Date column}
\item{...}{other non-specified parameters}
}
\description{
This graphic originally appeared
\href{http://stat-computing.org/dataexpo/2009/posters/wicklin-allison.pdf}{here}.
This function is included with the \code{makeR} package
to support the R-Bloggers demo. See \code{demo('makeR')}
for more information.
}
\author{
Paul Bleicher
}
| /man/calendarHeat.Rd | no_license | jbryer/makeR | R | false | false | 927 | rd | \name{calendarHeat}
\alias{calendarHeat}
\title{An R function to display time-series data as a calendar heatmap}
\usage{
calendarHeat(dates, values, ncolors = 99, color = "r2g",
varname = "Values", date.form = "\%Y-\%m-\%d", ...)
}
\arguments{
\item{dates}{vector of Dates}
\item{values}{vector of values}
\item{ncolors}{number of colors to use}
\item{color}{the color scheme to use. Currently supports
r2b, (red to blue), r2g (red to green), and w2b (white to
blue).}
\item{varname}{varaible names}
\item{date.form}{the format of the Date column}
\item{...}{other non-specified parameters}
}
\description{
This graphic originally appeared
\href{http://stat-computing.org/dataexpo/2009/posters/wicklin-allison.pdf}{here}.
This function is included with the \code{makeR} package
to support the R-Bloggers demo. See \code{demo('makeR')}
for more information.
}
\author{
Paul Bleicher
}
|
\name{countries}
\alias{countries}
\docType{data}
\title{Socioeconomic data for the most populous countries.}
\description{
Socioeconomic data for the most populous countries.
}
\usage{data(countries)}
\format{
A data frame with 42 observations on the following 7 variables.
\describe{
\item{Country}{name of the country.}
\item{Popul}{population.}
\item{PopDens}{population density.}
\item{GDPpp}{GDP per inhabitant.}
\item{LifeEx}{mean life expectation}
\item{InfMor}{infant mortality}
\item{Illit}{illiteracy rate}
}
}
\source{
CIA World Factbook \url{https://www.cia.gov/the-world-factbook/}
}
\examples{
data(countries)
summary(countries)
}
\keyword{datasets}
| /man/countries.Rd | no_license | cran/klaR | R | false | false | 731 | rd | \name{countries}
\alias{countries}
\docType{data}
\title{Socioeconomic data for the most populous countries.}
\description{
Socioeconomic data for the most populous countries.
}
\usage{data(countries)}
\format{
A data frame with 42 observations on the following 7 variables.
\describe{
\item{Country}{name of the country.}
\item{Popul}{population.}
\item{PopDens}{population density.}
\item{GDPpp}{GDP per inhabitant.}
\item{LifeEx}{mean life expectation}
\item{InfMor}{infant mortality}
\item{Illit}{illiteracy rate}
}
}
\source{
CIA World Factbook \url{https://www.cia.gov/the-world-factbook/}
}
\examples{
data(countries)
summary(countries)
}
\keyword{datasets}
|
install.packages("qdap")
install.packages("qdapTools")
library(qdapTools)
library(tm)
Source_dir <- c("d:/data")
Find_word <- c("management")
mylist <-c("here is the list")
files1 <-list.files(Source_dir,pattern=".docx")
# files1
#
# count1 <-length(table(files1))
# count1
#
# files1[1]
#
# name_file <- paste("d:/data/",files1[1],sep = "")
# name_file
#
# files1[11]
#
# read_docx(paste("d:/data/ Good Requirement.docx"))
# paste("d:/data/",files1[11],sep = "")
# sprintf("d:/data/",files1[11])
# data_1<- read_docx(paste("d:/data/",files1[11],sep = ""))
# data_1
#
count_occurence <- function (y) {
(text_docx <- read_docx(paste("d:/data/",y,sep = "")))
x <- length(grep("Find_word",text_docx))
return(x)
}
for (i in files1){
x<- count_occurence(i)
print(x)
print(i)
if (x!=0){ mylist[[length(mylist)+1]] <- i }
}
mylist
# if (length(count_occurence)==0){ mylist[[length(mylist)+1]] <- y }
#
data2 <-docx_data(files1[1])
# data2
#
count_occurence <- grep("management",data2)
length(count_occurence)
# str(count_occurence)
#if (length(count_occurence)==0){ mylist[[length(mylist)+1]] <- y }
lapply(files1,docx_data)
for(i in files1){ data1 <- cbind("d:/data/", i)
data1}
for(i in files1){ print(i)}
val <- LETTERS[1:4]
for ( i in val) {
print(i)
}
str(val)
str(files1) | /Finding the right document_1.R | no_license | karankkamra1987/R-Practice | R | false | false | 1,425 | r | install.packages("qdap")
install.packages("qdapTools")
library(qdapTools)
library(tm)
Source_dir <- c("d:/data")
Find_word <- c("management")
mylist <-c("here is the list")
files1 <-list.files(Source_dir,pattern=".docx")
# files1
#
# count1 <-length(table(files1))
# count1
#
# files1[1]
#
# name_file <- paste("d:/data/",files1[1],sep = "")
# name_file
#
# files1[11]
#
# read_docx(paste("d:/data/ Good Requirement.docx"))
# paste("d:/data/",files1[11],sep = "")
# sprintf("d:/data/",files1[11])
# data_1<- read_docx(paste("d:/data/",files1[11],sep = ""))
# data_1
#
count_occurence <- function (y) {
(text_docx <- read_docx(paste("d:/data/",y,sep = "")))
x <- length(grep("Find_word",text_docx))
return(x)
}
for (i in files1){
x<- count_occurence(i)
print(x)
print(i)
if (x!=0){ mylist[[length(mylist)+1]] <- i }
}
mylist
# if (length(count_occurence)==0){ mylist[[length(mylist)+1]] <- y }
#
data2 <-docx_data(files1[1])
# data2
#
count_occurence <- grep("management",data2)
length(count_occurence)
# str(count_occurence)
#if (length(count_occurence)==0){ mylist[[length(mylist)+1]] <- y }
lapply(files1,docx_data)
for(i in files1){ data1 <- cbind("d:/data/", i)
data1}
for(i in files1){ print(i)}
val <- LETTERS[1:4]
for ( i in val) {
print(i)
}
str(val)
str(files1) |
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/nn.R
\name{nn}
\alias{nn}
\title{Get names and class of all columns in a data frame}
\usage{
nn(df)
}
\arguments{
\item{df}{A data.frame.}
}
\value{
A data.frame with index and class.
}
\description{
Get names and class of all columns in a data frame in a friendly format.
}
\examples{
nn(iris)
}
\author{
Stephen Turner
}
\keyword{NA}
| /man/nn.Rd | no_license | gmaubach/Tmisc | R | false | true | 416 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/nn.R
\name{nn}
\alias{nn}
\title{Get names and class of all columns in a data frame}
\usage{
nn(df)
}
\arguments{
\item{df}{A data.frame.}
}
\value{
A data.frame with index and class.
}
\description{
Get names and class of all columns in a data frame in a friendly format.
}
\examples{
nn(iris)
}
\author{
Stephen Turner
}
\keyword{NA}
|
library(magick)
library(grid)
ggname = ggimage:::ggname
color_image = ggimage:::color_image
# color_image = function (img, color, alpha = NULL)
# {
# if (is.null(color))
# return(img)
# if (length(color) > 1) {
# stop("color should be a vector of length 1")
# }
# bitmap <- img[[1]]
# col <- col2rgb(color)
# bitmap[1, , ] <- as.raw(col[1])
# bitmap[2, , ] <- as.raw(col[2])
# bitmap[3, , ] <- as.raw(col[3])
# if (!is.null(alpha) && alpha != 1)
# browser()
# if(dim(bitmap)[1] == 3){
# bitmap[,4] = as.raw(255)
# }
# bitmap[4, , ] <- as.raw(as.integer(bitmap[4, , ]) * alpha)
# image_read(bitmap)
# }
##' geom layer for visualizing image files
##'
##'
##' @title geom_image.rect
##' @param mapping aes mapping
##' @param data data
##' @param stat stat
##' @param position position
##' @param inherit.aes logical, whether inherit aes from ggplot()
##' @param na.rm logical, whether remove NA values
##' @param by one of 'width' or 'height'
##' @param nudge_x horizontal adjustment to nudge image
##' @param ... additional parameters
##' @return geom layer
##' @importFrom ggplot2 layer
##' @export
##' @examples
##' library("ggplot2")
##' library("ggimage")
##' set.seed(2017-02-21)
##' d <- data.frame(x = rnorm(10),
##' y = rnorm(10),
##' image = sample(c("https://www.r-project.org/logo/Rlogo.png",
##' "https://jeroenooms.github.io/images/frink.png"),
##' size=10, replace = TRUE)
##' )
##' ggplot(d, aes(x, y)) + geom_image(aes(image=image))
##' @author guangchuang yu
geom_image.rect <- function(mapping=NULL, data=NULL, stat="identity",
position="identity", inherit.aes=TRUE,
na.rm=FALSE,
# by="width",
# nudge_x = 0,
...) {
# by <- match.arg(by, c("width", "height"))
layer(
data=data,
mapping=mapping,
geom=GeomImage.rect,
stat=stat,
position=position,
show.legend=NA,
inherit.aes=inherit.aes,
params = list(
na.rm = na.rm,
# by = by,
# nudge_x = nudge_x,
##angle = angle,
...),
check.aes = FALSE
)
}
##' @importFrom ggplot2 ggproto
##' @importFrom ggplot2 Geom
##' @importFrom ggplot2 aes
##' @importFrom ggplot2 draw_key_blank
##' @importFrom grid gTree
##' @importFrom grid gList
GeomImage.rect <- ggproto("GeomImage.rect", Geom,
setup_data = function(data, params) {
if (is.null(data$subset))
return(data)
data[which(data$subset),]
},
default_aes = aes(image=system.file("extdata/Rlogo.png", package="ggimage"),
#size=0.05,
colour = NULL, #angle = 0,
alpha=1),
draw_panel = function(data, panel_params, coord, by, na.rm=FALSE,
.fun = NULL, height, image_fun = NULL,
# hjust=0.5,
# nudge_x = 0, nudge_y = 0,
asp=1) {
# data$x <- data$x + nudge_x
# data$y <- data$y + nudge_y
data <- coord$transform(data, panel_params)
if (!is.null(.fun) && is.function(.fun))
data$image <- .fun(data$image)
groups <- split(data, factor(data$image))
imgs <- names(groups)
grobs <- lapply(seq_along(groups), function(i) {
d <- groups[[i]]
imageGrob.rect(d$xmin, d$xmax, d$ymin, d$ymax, imgs[i], #by,
# hjust,
d$colour, d$alpha, image_fun, #d$angle,
asp)
})
grobs <- do.call("c", grobs)
class(grobs) <- "gList"
ggname("geom_image.rect",
gTree(children = grobs))
},
non_missing_aes = c(#"size",
"image"),
required_aes = c("xmin", "xmax", "ymin", "ymax"),
draw_key = draw_key_image ## draw_key_blank ## need to write the `draw_key_image` function.
)
##' @importFrom magick image_read
##' @importFrom magick image_read_svg
##' @importFrom magick image_read_pdf
##' @importFrom magick image_transparent
##' @importFrom magick image_rotate
##' @importFrom grid rasterGrob
##' @importFrom grid viewport
##' @importFrom grDevices rgb
##' @importFrom grDevices col2rgb
##' @importFrom methods is
##' @importFrom tools file_ext
imageGrob.rect <- function(xmin, xmax, ymin, ymax, img, #by, hjust,
colour, alpha, image_fun, #angle,
asp=1) {
if (!is(img, "magick-image")) {
if (tools::file_ext(img) == "svg") {
img <- image_read_svg(img)
} else if (tools::file_ext(img) == "pdf") {
img <- image_read_pdf(img)
} else {
img <- image_read(img)
}
asp <- getAR2(img)/asp
}
unit <- "native"
width = xmax - xmin
height = ymax - ymin
# if (any(size == Inf)) {
# x <- 0.5
# y <- 0.5
# width <- 1
# height <- 1
# unit <- "npc"
# } else if (by == "width") {
# width <- size
# height <- size/asp
# } else {
# width <- size * asp
# height <- size
# }
#
# if (hjust == 0 || hjust == "left") {
# x <- x + width/2
# } else if (hjust == 1 || hjust == "right") {
# x <- x - width/2
# }
if (!is.null(image_fun)) {
img <- image_fun(img)
}
if (is.null(colour)) {
grobs <- list()
grobs[[1]] <- rasterGrob(x = xmin,
y = ymin,
just = c(0,0),
image = img,
default.units = unit,
height = height,
width = width,
interpolate = FALSE)
} else {
cimg <- lapply(seq_along(colour), function(i) {
color_image(img, colour[i], alpha[i])
})
grobs <- lapply(seq_along(xmin), function(i) {
img <- cimg[[i]]
# if (angle[i] != 0) {
# img <- image_rotate(img, angle[i])
# img <- image_transparent(img, "white")
# }
rasterGrob(x = xmin[i],
y = ymin[i],
just = c(0,0),
image = img,
default.units = unit,
height = height[i],
width = width[i],
interpolate = FALSE
## gp = gpar(rot = angle[i])
## vp = viewport(angle=angle[i])
)
})
}
return(grobs)
}
##' @importFrom magick image_info
getAR2 <- function(magick_image) {
info <- image_info(magick_image)
info$width/info$height
}
compute_just <- getFromNamespace("compute_just", "ggplot2") | /geom_image.rect.R | no_license | jrboyd/waldron | R | false | false | 7,889 | r | library(magick)
library(grid)
ggname = ggimage:::ggname
color_image = ggimage:::color_image
# color_image = function (img, color, alpha = NULL)
# {
# if (is.null(color))
# return(img)
# if (length(color) > 1) {
# stop("color should be a vector of length 1")
# }
# bitmap <- img[[1]]
# col <- col2rgb(color)
# bitmap[1, , ] <- as.raw(col[1])
# bitmap[2, , ] <- as.raw(col[2])
# bitmap[3, , ] <- as.raw(col[3])
# if (!is.null(alpha) && alpha != 1)
# browser()
# if(dim(bitmap)[1] == 3){
# bitmap[,4] = as.raw(255)
# }
# bitmap[4, , ] <- as.raw(as.integer(bitmap[4, , ]) * alpha)
# image_read(bitmap)
# }
##' geom layer for visualizing image files
##'
##'
##' @title geom_image.rect
##' @param mapping aes mapping
##' @param data data
##' @param stat stat
##' @param position position
##' @param inherit.aes logical, whether inherit aes from ggplot()
##' @param na.rm logical, whether remove NA values
##' @param by one of 'width' or 'height'
##' @param nudge_x horizontal adjustment to nudge image
##' @param ... additional parameters
##' @return geom layer
##' @importFrom ggplot2 layer
##' @export
##' @examples
##' library("ggplot2")
##' library("ggimage")
##' set.seed(2017-02-21)
##' d <- data.frame(x = rnorm(10),
##' y = rnorm(10),
##' image = sample(c("https://www.r-project.org/logo/Rlogo.png",
##' "https://jeroenooms.github.io/images/frink.png"),
##' size=10, replace = TRUE)
##' )
##' ggplot(d, aes(x, y)) + geom_image(aes(image=image))
##' @author guangchuang yu
geom_image.rect <- function(mapping=NULL, data=NULL, stat="identity",
position="identity", inherit.aes=TRUE,
na.rm=FALSE,
# by="width",
# nudge_x = 0,
...) {
# by <- match.arg(by, c("width", "height"))
layer(
data=data,
mapping=mapping,
geom=GeomImage.rect,
stat=stat,
position=position,
show.legend=NA,
inherit.aes=inherit.aes,
params = list(
na.rm = na.rm,
# by = by,
# nudge_x = nudge_x,
##angle = angle,
...),
check.aes = FALSE
)
}
##' @importFrom ggplot2 ggproto
##' @importFrom ggplot2 Geom
##' @importFrom ggplot2 aes
##' @importFrom ggplot2 draw_key_blank
##' @importFrom grid gTree
##' @importFrom grid gList
GeomImage.rect <- ggproto("GeomImage.rect", Geom,
setup_data = function(data, params) {
if (is.null(data$subset))
return(data)
data[which(data$subset),]
},
default_aes = aes(image=system.file("extdata/Rlogo.png", package="ggimage"),
#size=0.05,
colour = NULL, #angle = 0,
alpha=1),
draw_panel = function(data, panel_params, coord, by, na.rm=FALSE,
.fun = NULL, height, image_fun = NULL,
# hjust=0.5,
# nudge_x = 0, nudge_y = 0,
asp=1) {
# data$x <- data$x + nudge_x
# data$y <- data$y + nudge_y
data <- coord$transform(data, panel_params)
if (!is.null(.fun) && is.function(.fun))
data$image <- .fun(data$image)
groups <- split(data, factor(data$image))
imgs <- names(groups)
grobs <- lapply(seq_along(groups), function(i) {
d <- groups[[i]]
imageGrob.rect(d$xmin, d$xmax, d$ymin, d$ymax, imgs[i], #by,
# hjust,
d$colour, d$alpha, image_fun, #d$angle,
asp)
})
grobs <- do.call("c", grobs)
class(grobs) <- "gList"
ggname("geom_image.rect",
gTree(children = grobs))
},
non_missing_aes = c(#"size",
"image"),
required_aes = c("xmin", "xmax", "ymin", "ymax"),
draw_key = draw_key_image ## draw_key_blank ## need to write the `draw_key_image` function.
)
##' @importFrom magick image_read
##' @importFrom magick image_read_svg
##' @importFrom magick image_read_pdf
##' @importFrom magick image_transparent
##' @importFrom magick image_rotate
##' @importFrom grid rasterGrob
##' @importFrom grid viewport
##' @importFrom grDevices rgb
##' @importFrom grDevices col2rgb
##' @importFrom methods is
##' @importFrom tools file_ext
imageGrob.rect <- function(xmin, xmax, ymin, ymax, img, #by, hjust,
colour, alpha, image_fun, #angle,
asp=1) {
if (!is(img, "magick-image")) {
if (tools::file_ext(img) == "svg") {
img <- image_read_svg(img)
} else if (tools::file_ext(img) == "pdf") {
img <- image_read_pdf(img)
} else {
img <- image_read(img)
}
asp <- getAR2(img)/asp
}
unit <- "native"
width = xmax - xmin
height = ymax - ymin
# if (any(size == Inf)) {
# x <- 0.5
# y <- 0.5
# width <- 1
# height <- 1
# unit <- "npc"
# } else if (by == "width") {
# width <- size
# height <- size/asp
# } else {
# width <- size * asp
# height <- size
# }
#
# if (hjust == 0 || hjust == "left") {
# x <- x + width/2
# } else if (hjust == 1 || hjust == "right") {
# x <- x - width/2
# }
if (!is.null(image_fun)) {
img <- image_fun(img)
}
if (is.null(colour)) {
grobs <- list()
grobs[[1]] <- rasterGrob(x = xmin,
y = ymin,
just = c(0,0),
image = img,
default.units = unit,
height = height,
width = width,
interpolate = FALSE)
} else {
cimg <- lapply(seq_along(colour), function(i) {
color_image(img, colour[i], alpha[i])
})
grobs <- lapply(seq_along(xmin), function(i) {
img <- cimg[[i]]
# if (angle[i] != 0) {
# img <- image_rotate(img, angle[i])
# img <- image_transparent(img, "white")
# }
rasterGrob(x = xmin[i],
y = ymin[i],
just = c(0,0),
image = img,
default.units = unit,
height = height[i],
width = width[i],
interpolate = FALSE
## gp = gpar(rot = angle[i])
## vp = viewport(angle=angle[i])
)
})
}
return(grobs)
}
##' @importFrom magick image_info
getAR2 <- function(magick_image) {
info <- image_info(magick_image)
info$width/info$height
}
compute_just <- getFromNamespace("compute_just", "ggplot2") |
regwqComp <- function (formula, data, alpha=.05)
{
# this is a reworked version of regwq.R from the mutoss package
# rounding is removed
# output is reordered in a standard order
# sqrt(2) factor removed from test statistic (and adjusted elsewhere)
#
# important - the adjusted p values are compared to the adjusted alphas
# it is NOT enough to compare adjusted p to the original alpha
#
# no guarantees on the procedure itself as that is preserved from mutoss
if (missing(data)) {
dat <- model.frame(formula)
} else {
dat <- model.frame(formula, data)
}
if (ncol(dat) != 2) {
stop("Specify one response and only one class variable in the formula")
}
if (is.numeric(dat[, 1]) == FALSE) {
stop("Response variable must be numeric")
}
response <- dat[, 1]
group <- as.factor(dat[, 2])
fl <- levels(group)
a <- nlevels(group)
N <- length(response)
samples <- split(response, group)
n <- sapply(samples, "length")
mm <- sapply(samples, "mean")
vv <- sapply(samples, "var")
MSE <- sum((n - 1) * vv)/(N - a)
df <- N - a
nc <- a * (a - 1)/2
order.h1 <- data.frame(Sample = fl, SampleNum = 1:a, Size = n, Means = mm,
Variance = vv)
ordered <- order.h1[order(order.h1$Means, decreasing = FALSE),
]
rownames(ordered) <- 1:a
i <- 1:(a - 1)
h1 <- list()
for (s in 1:(a - 1)) {
h1[[s]] <- i[1:s]
}
vi <- unlist(h1)
j <- a:2
h2 <- list()
for (s in 1:(a - 1)) {
h2[[s]] <- j[s:1]
}
vj <- unlist(h2)
h3 <- list()
h4 <- list()
for (s in 1:(a - 1)) {
h3[[s]] <- rep(j[s], s)
h4[[s]] <- rep(i[s], s)
}
Nmean <- unlist(h3)
Step <- unlist(h4)
mean.difference <- sapply(1:nc, function(arg) {
i <- vi[arg]
j <- vj[arg]
(ordered$Means[j] - ordered$Means[i])
})
T <- sapply(1:nc, function(arg) {
i <- vi[arg]
j <- vj[arg]
(ordered$Means[j] - ordered$Means[i])/sqrt(MSE * (1/ordered$Size[i] +
1/ordered$Size[j]))
})
pvalues <- ptukey(T*sqrt(2), Nmean, df, lower.tail = FALSE)
alpha.level <- 1 - (1 - alpha)^(Nmean/a)
level1 <- (Nmean == a)
level2 <- (Nmean == a - 1)
level3 <- level1 + level2
alpha.level[level3 == 1] <- alpha
quantiles <- qtukey(1 - alpha.level, Nmean, df)
for (h in 1:(nc - 1)) {
if (quantiles[h + 1] >= quantiles[h]) {
quantiles[h + 1] <- quantiles[h]
}
}
Rejected1 <- (pvalues < alpha.level)
names.ordered <- sapply(1:nc, function(arg) {
i <- vi[arg]
j <- vj[arg]
paste(ordered$Sample[j], "-", ordered$Sample[i], sep = "")
})
for (s in 1:nc) {
if (Rejected1[s] == FALSE) {
Under1 <- (vj[s] >= vj)
Under2 <- (vi[s] <= vi)
Under3 <- Under1 * Under2
Under4 <- which(Under3 == 1)
Rejected1[Under4] <- FALSE
}
}
# add code here to put rows in standard order and flip diff and T if necessary
# first build a matrix that will contain the desired row numbers
rowOrderMat <- matrix(0,a,a)
rowOrderMat[lower.tri(rowOrderMat)]<-1:nc
rownames(rowOrderMat) <- fl
colnames(rowOrderMat) <- fl
rowOrderVec <- numeric(nc)
signVec <- rep(1,nc)
for (s in 1:nc){
i <- vi[s]
j <- vj[s]
si <- ordered$SampleNum[i]
sj <- ordered$SampleNum[j]
if (si<sj){
rowOrderVec[s] <- rowOrderMat[sj,si]
} else{
rowOrderVec[s] <- rowOrderMat[si,sj]
names.ordered[s] <- paste(ordered$Sample[i], "-", ordered$Sample[j], sep = "")
mean.difference[s] <- -mean.difference[s]
T[s] <- -T[s]
}
}
ind <- order(rowOrderVec)
pvalues <- pvalues[ind]
mean.difference <- mean.difference[ind]
T <- T[ind]
names.ordered <- names.ordered[ind]
Rejected1 <- Rejected1[ind]
alpha.level <- alpha.level[ind]
pv <- 2*pt(-abs(T),df=N-a)
comp.matrix <- cbind(mean.difference,T,pv,pvalues,alpha.level,as.numeric(Rejected1))
dimnames(comp.matrix) <- list(names.ordered,c("diff","t","p","p adj","alpha adj","rej H_0"))
return(comp.matrix)
} | /R/regwqComp.R | no_license | DataScienceUWL/DS705data | R | false | false | 4,060 | r | regwqComp <- function (formula, data, alpha=.05)
{
# this is a reworked version of regwq.R from the mutoss package
# rounding is removed
# output is reordered in a standard order
# sqrt(2) factor removed from test statistic (and adjusted elsewhere)
#
# important - the adjusted p values are compared to the adjusted alphas
# it is NOT enough to compare adjusted p to the original alpha
#
# no guarantees on the procedure itself as that is preserved from mutoss
if (missing(data)) {
dat <- model.frame(formula)
} else {
dat <- model.frame(formula, data)
}
if (ncol(dat) != 2) {
stop("Specify one response and only one class variable in the formula")
}
if (is.numeric(dat[, 1]) == FALSE) {
stop("Response variable must be numeric")
}
response <- dat[, 1]
group <- as.factor(dat[, 2])
fl <- levels(group)
a <- nlevels(group)
N <- length(response)
samples <- split(response, group)
n <- sapply(samples, "length")
mm <- sapply(samples, "mean")
vv <- sapply(samples, "var")
MSE <- sum((n - 1) * vv)/(N - a)
df <- N - a
nc <- a * (a - 1)/2
order.h1 <- data.frame(Sample = fl, SampleNum = 1:a, Size = n, Means = mm,
Variance = vv)
ordered <- order.h1[order(order.h1$Means, decreasing = FALSE),
]
rownames(ordered) <- 1:a
i <- 1:(a - 1)
h1 <- list()
for (s in 1:(a - 1)) {
h1[[s]] <- i[1:s]
}
vi <- unlist(h1)
j <- a:2
h2 <- list()
for (s in 1:(a - 1)) {
h2[[s]] <- j[s:1]
}
vj <- unlist(h2)
h3 <- list()
h4 <- list()
for (s in 1:(a - 1)) {
h3[[s]] <- rep(j[s], s)
h4[[s]] <- rep(i[s], s)
}
Nmean <- unlist(h3)
Step <- unlist(h4)
mean.difference <- sapply(1:nc, function(arg) {
i <- vi[arg]
j <- vj[arg]
(ordered$Means[j] - ordered$Means[i])
})
T <- sapply(1:nc, function(arg) {
i <- vi[arg]
j <- vj[arg]
(ordered$Means[j] - ordered$Means[i])/sqrt(MSE * (1/ordered$Size[i] +
1/ordered$Size[j]))
})
pvalues <- ptukey(T*sqrt(2), Nmean, df, lower.tail = FALSE)
alpha.level <- 1 - (1 - alpha)^(Nmean/a)
level1 <- (Nmean == a)
level2 <- (Nmean == a - 1)
level3 <- level1 + level2
alpha.level[level3 == 1] <- alpha
quantiles <- qtukey(1 - alpha.level, Nmean, df)
for (h in 1:(nc - 1)) {
if (quantiles[h + 1] >= quantiles[h]) {
quantiles[h + 1] <- quantiles[h]
}
}
Rejected1 <- (pvalues < alpha.level)
names.ordered <- sapply(1:nc, function(arg) {
i <- vi[arg]
j <- vj[arg]
paste(ordered$Sample[j], "-", ordered$Sample[i], sep = "")
})
for (s in 1:nc) {
if (Rejected1[s] == FALSE) {
Under1 <- (vj[s] >= vj)
Under2 <- (vi[s] <= vi)
Under3 <- Under1 * Under2
Under4 <- which(Under3 == 1)
Rejected1[Under4] <- FALSE
}
}
# add code here to put rows in standard order and flip diff and T if necessary
# first build a matrix that will contain the desired row numbers
rowOrderMat <- matrix(0,a,a)
rowOrderMat[lower.tri(rowOrderMat)]<-1:nc
rownames(rowOrderMat) <- fl
colnames(rowOrderMat) <- fl
rowOrderVec <- numeric(nc)
signVec <- rep(1,nc)
for (s in 1:nc){
i <- vi[s]
j <- vj[s]
si <- ordered$SampleNum[i]
sj <- ordered$SampleNum[j]
if (si<sj){
rowOrderVec[s] <- rowOrderMat[sj,si]
} else{
rowOrderVec[s] <- rowOrderMat[si,sj]
names.ordered[s] <- paste(ordered$Sample[i], "-", ordered$Sample[j], sep = "")
mean.difference[s] <- -mean.difference[s]
T[s] <- -T[s]
}
}
ind <- order(rowOrderVec)
pvalues <- pvalues[ind]
mean.difference <- mean.difference[ind]
T <- T[ind]
names.ordered <- names.ordered[ind]
Rejected1 <- Rejected1[ind]
alpha.level <- alpha.level[ind]
pv <- 2*pt(-abs(T),df=N-a)
comp.matrix <- cbind(mean.difference,T,pv,pvalues,alpha.level,as.numeric(Rejected1))
dimnames(comp.matrix) <- list(names.ordered,c("diff","t","p","p adj","alpha adj","rej H_0"))
return(comp.matrix)
} |
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/CINNA.R
\name{print.visualize.heatmap}
\alias{print.visualize.heatmap}
\title{Print the heatmap plot of centrality measures}
\usage{
\method{print}{visualize.heatmap}(x, scale = TRUE, file = NULL)
}
\arguments{
\item{x}{a list indicating calculated centrality measures}
\item{scale}{Whether the centrality values should be scaled or not(default=TRUE)}
\item{file}{A character string naming the .pdf file to print into. If NULL
the result would be printed to the exist directory.(default=NULL)}
}
\value{
The resulted plot of \code{ \link[CINNA]{visualize_heatmap}}function will be saved in the given directory.
}
\description{
This function prints the heatmap plot
}
\author{
Minoo Ashtiani, Mohieddin Jafari
}
| /man/print.visualize.heatmap.Rd | no_license | jafarilab/CINNA | R | false | true | 791 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/CINNA.R
\name{print.visualize.heatmap}
\alias{print.visualize.heatmap}
\title{Print the heatmap plot of centrality measures}
\usage{
\method{print}{visualize.heatmap}(x, scale = TRUE, file = NULL)
}
\arguments{
\item{x}{a list indicating calculated centrality measures}
\item{scale}{Whether the centrality values should be scaled or not(default=TRUE)}
\item{file}{A character string naming the .pdf file to print into. If NULL
the result would be printed to the exist directory.(default=NULL)}
}
\value{
The resulted plot of \code{ \link[CINNA]{visualize_heatmap}}function will be saved in the given directory.
}
\description{
This function prints the heatmap plot
}
\author{
Minoo Ashtiani, Mohieddin Jafari
}
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/methods-SingleBatchModel.R
\name{MarginalModelList}
\alias{MarginalModelList}
\title{Constructor for list of single-batch models}
\usage{
MarginalModelList(data = numeric(), k = numeric(),
mcmc.params = McmcParams(), ...)
}
\arguments{
\item{data}{numeric vector of average log R ratios}
\item{k}{numeric vector indicating the number of mixture components for each model}
\item{mcmc.params}{an object of class \code{McmcParams}}
\item{...}{additional arguments passed to \code{Hyperparameters}}
}
\value{
a list. Each element of the list is a \code{BatchModel}
}
\description{
An object of class MarginalModel is constructed for each k, creating a list of
MarginalModels.
}
\examples{
mlist <- MarginalModelList(data=y(MarginalModelExample), k=1:4)
mcmcParams(mlist) <- McmcParams(iter=1, burnin=1, nStarts=0)
mlist2 <- posteriorSimulation(mlist)
}
\seealso{
\code{\link{MarginalModel}} \code{\link{BatchModelList}}
}
| /man/MarginalModelList.Rd | no_license | muschellij2/CNPBayes | R | false | true | 1,001 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/methods-SingleBatchModel.R
\name{MarginalModelList}
\alias{MarginalModelList}
\title{Constructor for list of single-batch models}
\usage{
MarginalModelList(data = numeric(), k = numeric(),
mcmc.params = McmcParams(), ...)
}
\arguments{
\item{data}{numeric vector of average log R ratios}
\item{k}{numeric vector indicating the number of mixture components for each model}
\item{mcmc.params}{an object of class \code{McmcParams}}
\item{...}{additional arguments passed to \code{Hyperparameters}}
}
\value{
a list. Each element of the list is a \code{BatchModel}
}
\description{
An object of class MarginalModel is constructed for each k, creating a list of
MarginalModels.
}
\examples{
mlist <- MarginalModelList(data=y(MarginalModelExample), k=1:4)
mcmcParams(mlist) <- McmcParams(iter=1, burnin=1, nStarts=0)
mlist2 <- posteriorSimulation(mlist)
}
\seealso{
\code{\link{MarginalModel}} \code{\link{BatchModelList}}
}
|
#' Plots an FFTrees object.
#'
#' @description Plots an FFTrees object created by the FFTrees() function.
#' @param x A FFTrees object created from \code{"FFTrees()"}
#' @param data One of two strings 'train' or 'test'. In this case, the corresponding dataset in the x object will be used.
#' @param what string. What should be plotted? \code{'tree'} (the default) shows one tree (specified by \code{'tree'}). \code{'cues'} shows the marginal accuracy of cues in an ROC space, \code{"roc"} shows an roc curve of the tree(s)
#' @param tree integer. An integer indicating which tree to plot (only valid when the tree argument is non-empty). To plot the best training (or test) tree with respect to the \code{goal} specified during FFT construction, use "best.train" or "best.test"
#' @param main character. The main plot label.
#' @param hlines logical. Should horizontal panel separation lines be shown?
#' @param cue.labels character. An optional string of labels for the cues / nodes.
#' @param decision.labels character. A string vector of length 2 indicating the content-specific name for noise and signal cases.
#' @param cue.cex numeric. The size of the cue labels.
#' @param threshold.cex numeric. The size of the threshold labels.
#' @param decision.cex numeric. The size of the decision labels.
#' @param comp logical. Should the performance of competitive algorithms (e.g.; logistic regression, random forests etc.) be shown in the ROC plot (if available?)
#' @param stats logical. Should statistical information be plotted? If \code{FALSE}, then only the tree (without any reference to statistics) will be plotted.
#' @param show.header,show.tree,show.confusion,show.levels,show.roc,show.icons,show.iconguide logical. Logical arguments indicating which specific elements of the plot to show.
#' @param label.tree,label.performance string. Optional arguments to define lables for the tree and performance section(s).
#' @param n.per.icon Number of cases per icon
#' @param which.tree deprecated argument, only for backwards compatibility, use \code{"tree"} instead.
#' @param level.type string. How should bottom levels be drawn? Can be \code{"bar"} or \code{"line"}
#' @param decision.names deprecated arguments.
#' @param ... Currently ignored.
#' @importFrom stats anova predict formula model.frame
#' @importFrom graphics text points abline legend mtext segments rect arrows axis par layout plot
#' @importFrom grDevices gray col2rgb rgb
#' @importFrom yarrr transparent piratepal
#' @export
#' @examples
#'
#' # Create FFTrees of the heart disease data
#' heart.fft <- FFTrees(formula = diagnosis ~.,
#' data = heartdisease)
#'
#' # Visualise the tree
#' plot(heart.fft,
#' main = "Heart Disease Diagnosis",
#' decision.labels = c("Absent", "Present"))
#'
#'
#' # See the vignette for more details
#' vignette("FFTrees_plot", package = "FFTrees")
#'
#'
#'
plot.FFTrees <- function(
x = NULL,
data = "train",
what = 'tree',
tree = "best.train",
main = NULL,
hlines = TRUE,
cue.labels = NULL,
decision.labels = NULL,
cue.cex = NULL,
threshold.cex = NULL,
decision.cex = 1,
comp = TRUE,
stats = TRUE,
show.header = NULL,
show.tree = NULL,
show.confusion = NULL,
show.levels = NULL,
show.roc = NULL,
show.icons = NULL,
show.iconguide = NULL,
label.tree = NULL,
label.performance = NULL,
n.per.icon = NULL,
which.tree = NULL,
level.type = "bar",
decision.names = NULL,
...
) {
#
# # #
# data = "train"
# what = 'tree'
# tree = "best.train"
# main = NULL
# hlines = TRUE
# cue.labels = NULL
# decision.labels = NULL
# cue.cex = NULL
# threshold.cex = NULL
# decision.cex = 1
# comp = TRUE
# stats = TRUE
# show.header = NULL
# show.tree = NULL
# show.confusion = NULL
# show.levels = NULL
# show.roc = NULL
# show.icons = NULL
# show.iconguide = NULL
# label.tree = NULL
# label.performance = NULL
# n.per.icon = NULL
# which.tree = NULL
# level.type = "bar"
# decision.names = NULL
# Check for invalid or missing arguments
# Input validation
{
if(what %in% c("cues", "tree", "roc") == FALSE) {
stop("what must be either 'cues', 'tree', or 'roc'")
}
if(is.null(decision.names) == FALSE) {
message("decision.names is deprecated, use decision.lables instead")
decision.labels <- decision.names
}
}
# If what == cues, then send inputs to showcues()
if(what == 'cues') {showcues(x = x, data = data, main = main)}
if(what != 'cues') {
# Determine layout
{
if(what == "tree") {
if(stats == TRUE) {
if(is.null(show.header)) {show.header <- TRUE}
if(is.null(show.tree)) {show.tree <- TRUE}
if(is.null(show.confusion)) {show.confusion <- TRUE}
if(is.null(show.levels)) {show.levels <- TRUE}
if(is.null(show.roc)) {show.roc <- TRUE}
if(is.null(show.icons)) {show.icons <- TRUE}
if(is.null(show.iconguide)) {show.iconguide <- TRUE}
}
if(stats == FALSE) {
if(is.null(show.header)) {show.header <- FALSE}
if(is.null(show.tree)) {show.tree <- TRUE}
if(is.null(show.confusion)) {show.confusion <- FALSE}
if(is.null(show.levels)) {show.levels <- FALSE}
if(is.null(show.roc)) {show.roc <- FALSE}
if(is.null(show.icons)) {show.icons <- FALSE}
if(is.null(show.iconguide)) {show.iconguide <- FALSE}
}
}
if(what == "roc") {
show.header <- FALSE
show.tree <- FALSE
show.confusion <- FALSE
show.levels <- FALSE
show.roc <- TRUE
show.icons <- FALSE
show.top <- FALSE
}
# Top, middle, bottom
if(show.header & show.tree & (show.confusion | show.levels | show.roc)) {
show.top <- TRUE
show.middle <- TRUE
show.bottom <- TRUE
layout(matrix(1:3, nrow = 3, ncol = 1),
widths = c(6),
heights = c(1.2, 3, 1.8))
}
# Top and middle
if(show.header & show.tree & (show.confusion == FALSE & show.levels == FALSE & show.roc == FALSE)) {
show.top <- TRUE
show.middle <- TRUE
show.bottom <- FALSE
layout(matrix(1:2, nrow = 2, ncol = 1),
widths = c(6),
heights = c(1.2, 3))
}
# Middle and bottom
if(show.header == FALSE & show.tree & (show.confusion | show.levels | show.roc)) {
show.top <- FALSE
show.middle <- TRUE
show.bottom <- TRUE
layout(matrix(1:2, nrow = 2, ncol = 1),
widths = c(6),
heights = c(3, 1.8))
}
# Middle
if(show.header == FALSE & show.tree & (show.confusion == FALSE & show.levels == FALSE & show.roc == FALSE)) {
show.top <- FALSE
show.middle <- TRUE
show.bottom <- FALSE
layout(matrix(1:1, nrow = 1, ncol = 1),
widths = c(6),
heights = c(3))
}
# Bottom
if(show.header == FALSE & show.tree == FALSE) {
show.top <- FALSE
show.middle <- FALSE
show.bottom <- TRUE
nplots <- show.confusion + show.levels + show.roc
layout(matrix(1:nplots, nrow = 1, ncol = nplots),
widths = c(3 * nplots),
heights = c(3))
}
}
# -------------------------
# Setup data
# --------------------------
{
# Extract important parameters from x
goal <- x$params$goal
if(is.null(decision.labels)) {
if(("decision.labels" %in% names(x$params))) {
decision.labels <- x$params$decision.labels
} else {decision.labels <- c(0, 1)}
}
if(is.null(main)) {
if(("main" %in% names(x$params))) {
if(is.null(x$params$main)) {
if(show.header) {
main <- "Data"
} else {main <- ""}
} else {
main <- x$params$main
}
} else {
if(class(data) == "character") {
if(data == "train") {main <- "Data (Training)"}
if(data == "test") {main <- "Data (Testing)"}
}
if(class(data) == "data.frame") {main <- "Test Data"}
}
}
# Check for problems and depreciated arguments
{
if(is.null(which.tree) == FALSE) {
message("The which.tree argument is depreciated and is now just called tree. Please use tree from now on to avoid this message.")
tree <- which.tree
}
if(class(x) != "FFTrees") {
stop("You did not include a valid FFTrees class object or specify the tree directly with level.names, level.classes (etc.). Either create a valid FFTrees object with FFTrees() or specify the tree directly.")
}
if(tree == "best.test" & is.null(x$tree.stats$test)) {
print("You wanted to plot the best test tree (tree = 'best.test') but there were no test data, I'll plot the best training tree instead")
tree <- "best.train"
}
if(is.numeric(tree) & (tree %in% 1:x$trees$n) == F) {
stop(paste("You asked for a tree that does not exist. This object has", x$trees$n, "trees"))
}
if(class(data) == "character") {
if(data == "test" & is.null(x$trees$results$test$stats)) {stop("You asked to plot the test data but there are no test data in the FFTrees object")}
}
}
# DEFINE PLOTTING TREE
if(tree == "best.train") {
tree <- x$trees$best$train
}
if(tree == "best.test") {
tree <- x$trees$best$test
}
# DEFINE CRITICAL OBJECTS
lr.stats <- NULL
cart.stats <- NULL
rf.stats <- NULL
svm.stats <- NULL
decision.v <- x$trees$results[[data]]$decisions[,tree]
tree.stats <- x$trees$results[[data]]$stats
level.stats <- x$trees$results[[data]]$level_stats[x$trees$results[[data]]$level_stats$tree == tree,]
if(comp == TRUE) {
if(is.null(x$comp$lr$results) == FALSE) {
lr.stats <- data.frame("sens" = x$comp$lr$results[[data]]$sens,
"spec" = x$comp$lr$results[[data]]$spec)
}
if(is.null(x$comp$cart$results) == FALSE) {
lcart.stats <- data.frame("sens" = x$comp$cart$results[[data]]$sens,
"spec" = x$comp$cart$results[[data]]$spec)
}
if(is.null(x$comp$rf$results) == FALSE) {
rf.stats <- data.frame("sens" = x$comp$rf$results[[data]]$sens,
"spec" = x$comp$rf$results[[data]]$spec)
}
if(is.null(x$comp$svm$results) == FALSE) {
svm.stats <- data.frame("sens" = x$comp$svm$results[[data]]$sens,
"spec" = x$comp$svm$results[[data]]$spec)
}
}
n.exemplars <- nrow(x$data[[data]])
n.pos <- sum(x$data[[data]][[x$metadata$criterion_name]])
n.neg <- sum(x$data[[data]][[x$metadata$criterion_name]] == FALSE)
mcu <- x$trees$results[[data]]$stats$mcu[tree]
crit.br <- mean(x$data[[data]][[x$metadata$criterion_name]])
final.stats <- tree.stats[tree,]
# ADD LEVEL STATISTICS
n.levels <- nrow(level.stats)
# Add marginal classification statistics to level.stats
level.stats[c("hi.m", "mi.m", "fa.m", "cr.m")] <- NA
for(i in 1:n.levels) {
if(i == 1) {
level.stats[1, c("hi.m", "mi.m", "fa.m", "cr.m")] <- level.stats[1, c("hi", "mi", "fa", "cr")]
}
if(i > 1) {
level.stats[i, c("hi.m", "mi.m", "fa.m", "cr.m")] <- level.stats[i, c("hi", "mi", "fa", "cr")] - level.stats[i - 1, c("hi", "mi", "fa", "cr")]
}
}
}
# -------------------------
# Define plotting parameters
# --------------------------
{
{
# Panels
panel.line.lwd <- 1
panel.line.col <- gray(0)
panel.line.lty <- 1
# Some general parameters
ball.col = c(gray(0), gray(0))
ball.bg = c(gray(1), gray(1))
ball.pch = c(21, 24)
ball.cex = c(1, 1)
error.col <- "red"
correct.col <- "green"
max.label.length <- 100
def.par <- par(no.readonly = TRUE)
ball.box.width <- 10
label.box.height <- 2
label.box.width <- 5
# Define cue labels
{
if(is.null(cue.labels)) {
cue.labels <- level.stats$cue
}
# Trim labels
cue.labels <- strtrim(cue.labels, max.label.length)
}
# Node Segments
segment.lty <- 1
segment.lwd <- 1
continue.segment.lwd <- 1
continue.segment.lty <- 1
exit.segment.lwd <- 1
exit.segment.lty <- 1
decision.node.cex <- 4
exit.node.cex <- 4
panel.title.cex <- 2
# Classification table
classtable.lwd <- 1
# ROC
roc.lwd <- 1
roc.border.col <- gray(0)
# Label sizes
{
# Set cue label size
if(is.null(cue.cex)) {
cue.cex <- c(1.5, 1.5, 1.25, 1, 1, 1)
} else {
if(length(cue.cex) < 6) {cue.cex <- rep(cue.cex, length.out = 6)}
}
# Set break label size
if(is.null(threshold.cex)) {
threshold.cex <- c(1.5, 1.5, 1.25, 1, 1, 1)
} else {
if(length(threshold.cex) < 6) {threshold.cex <- rep(threshold.cex, length.out = 6)}
}
}
if(show.top & show.middle & show.bottom) {
plotting.parameters.df <- data.frame(
n.levels = 1:6,
plot.height = c(10, 12, 15, 19, 23, 27),
plot.width = c(14, 16, 20, 24, 28, 34),
label.box.text.cex = cue.cex,
break.label.cex = threshold.cex
)
} else if (show.top == FALSE & show.middle & show.bottom == FALSE) {
plotting.parameters.df <- data.frame(
n.levels = 1:6,
plot.height = c(10, 12, 15, 19, 23, 25),
plot.width = c(14, 16, 20, 24, 28, 32) * 1,
label.box.text.cex = cue.cex,
break.label.cex = threshold.cex
)
} else {
plotting.parameters.df <- data.frame(
n.levels = 1:6,
plot.height = c(10, 12, 15, 19, 23, 25),
plot.width = c(14, 16, 20, 24, 28, 32) * 1,
label.box.text.cex = cue.cex,
break.label.cex = threshold.cex
)
}
if(n.levels < 6) {
label.box.text.cex <- plotting.parameters.df$label.box.text.cex[n.levels]
break.label.cex <- plotting.parameters.df$break.label.cex[n.levels]
plot.height <- plotting.parameters.df$plot.height[n.levels]
plot.width <- plotting.parameters.df$plot.width[n.levels]
}
if(n.levels >= 6) {
label.box.text.cex <- plotting.parameters.df$label.box.text.cex[6]
break.label.cex <- plotting.parameters.df$break.label.cex[6]
plot.height <- plotting.parameters.df$plot.height[6]
plot.width <- plotting.parameters.df$plot.width[6]
}
# Colors
exit.node.bg <- "white"
error.colfun <- circlize::colorRamp2(c(0, 50, 100),
colors = c("white", "red", "black"))
correct.colfun <- circlize::colorRamp2(c(0, 50, 100),
colors = c("white", "green", "black"))
error.bg <- yarrr::transparent(error.colfun(35), .2)
error.border <- yarrr::transparent(error.colfun(65), .1)
correct.bg <- yarrr::transparent(correct.colfun(35), .2)
correct.border <- yarrr::transparent(correct.colfun(65), .1)
max.cex <- 6
min.cex <- 1
exit.node.pch <- 21
decision.node.pch <- NA_integer_
# balls
ball.loc <- "variable"
if(n.levels == 3) {ball.box.width <- 14}
if(n.levels == 4) {ball.box.width <- 18}
ball.box.height <- 2.5
ball.box.horiz.shift <- 10
ball.box.vert.shift <- -1
ball.box.max.shift.p <- .9
ball.box.min.shift.p <- .4
ball.box.fixed.x.shift <- c(ball.box.min.shift.p * plot.width, ball.box.max.shift.p * plot.width)
# Determine N per ball
if(is.null(n.per.icon)) {
max.n.side <- max(c(n.pos, n.neg))
i <- max.n.side / c(1, 5, 10^(1:10))
i[i > 50] <- 0
n.per.icon <- c(1, 5, 10^(1:10))[which(i == max(i))]
}
noise.ball.pch <- ball.pch[1]
signal.ball.pch <- ball.pch[2]
noise.ball.col <- ball.col[1]
signal.ball.col <- ball.col[2]
noise.ball.bg <- ball.bg[1]
signal.ball.bg <- ball.bg[2]
# arrows
arrow.lty <- 1
arrow.lwd <- 1
arrow.length <- 2.5
arrow.head.length <- .08
arrow.col <- gray(0)
# Final stats
spec.circle.x <- .4
dprime.circle.x <- .5
sens.circle.x <- .6
stat.circle.y <- .3
sens.circle.col <- "green"
spec.circle.col <- "red"
dprime.circle.col <- "blue"
stat.outer.circle.col <- gray(.5)
}
# add.balls.fun() adds balls to the plot
add.balls.fun <- function(x.lim = c(-10, 0),
y.lim = c(-2, 0),
n.vec = c(20, 10),
pch.vec = c(21, 21),
ball.cex = 1,
bg.vec = ball.bg,
col.vec = ball.col,
ball.lwd = .7,
freq.text = TRUE,
freq.text.cex = 1.2,
upper.text = "",
upper.text.cex = 1,
upper.text.adj = 0,
rev.order = F,
box.col = NULL,
box.bg = NULL,
n.per.icon = NULL
) {
# Add box
if(is.null(box.col) == FALSE | is.null(box.bg) == FALSE) {
rect(x.lim[1],
y.lim[1],
x.lim[2],
y.lim[2],
col = box.bg,
border = box.col)
}
# add upper text
text(mean(x.lim), y.lim[2] + upper.text.adj,
label = upper.text, cex = upper.text.cex
)
a.n <- n.vec[1]
b.n <- n.vec[2]
a.p <- n.vec[1] / sum(n.vec)
box.x.center <- sum(x.lim) / 2
box.y.center <- sum(y.lim) / 2
box.x.width <- x.lim[2] - x.lim[1]
# Determine cases per ball
if(is.null(n.per.icon)) {
max.n.side <- max(c(a.n, b.n))
i <- max.n.side / c(1, 5, 10, 50, 100, 1000, 10000)
i[i > 50] <- 0
n.per.icon <- c(1, 5, 10, 50, 100, 1000, 10000)[which(i == max(i))]
}
# Determine general ball locations
a.balls <- ceiling(a.n / n.per.icon)
b.balls <- ceiling(b.n / n.per.icon)
n.balls <- a.balls + b.balls
a.ball.x.loc <- 0
a.ball.y.loc <- 0
b.ball.x.loc <- 0
b.ball.y.loc <- 0
if(a.balls > 0) {
a.ball.x.loc <- rep(-1:-10, each = 5, length.out = 50)[1:a.balls]
a.ball.y.loc <- rep(1:5, times = 10, length.out = 50)[1:a.balls]
a.ball.x.loc <- a.ball.x.loc * (x.lim[2] - box.x.center) / 10 + box.x.center
a.ball.y.loc <- a.ball.y.loc * (y.lim[2] - y.lim[1]) / 5 + y.lim[1]
}
if(b.balls > 0) {
b.ball.x.loc <- rep(1:10, each = 5, length.out = 50)[1:b.balls]
b.ball.y.loc <- rep(1:5, times = 10, length.out = 50)[1:b.balls]
b.ball.x.loc <- b.ball.x.loc * (x.lim[2] - box.x.center) / 10 + box.x.center
b.ball.y.loc <- b.ball.y.loc * (y.lim[2] - y.lim[1]) / 5 + y.lim[1]
}
# if(rev.order) {
#
# x <- b.ball.x.loc
# y <- b.ball.y.loc
#
# b.ball.x.loc <- a.x.loc
# b.ball.y.loc <- a.y.loc
#
# a.ball.x.loc <- x
# a.ball.y.loc <- y
#
# }
# Add frequency text
if(freq.text) {
text(box.x.center, y.lim[1] - 1 * (y.lim[2] - y.lim[1]) / 5, prettyNum(b.n, big.mark = ","), pos = 4, cex = freq.text.cex)
text(box.x.center, y.lim[1] - 1 * (y.lim[2] - y.lim[1]) / 5, prettyNum(a.n, big.mark = ","), pos = 2, cex = freq.text.cex)
}
# Draw balls
# Noise
suppressWarnings(if(a.balls > 0) {
points(x = a.ball.x.loc,
y = a.ball.y.loc,
pch = pch.vec[1],
bg = bg.vec[1],
col = col.vec[1],
cex = ball.cex,
lwd = ball.lwd)
}
)
# Signal
suppressWarnings(if(b.balls > 0) {
points(x = b.ball.x.loc,
y = b.ball.y.loc,
pch = pch.vec[2],
bg = bg.vec[2],
col = col.vec[2],
cex = ball.cex,
lwd = ball.lwd
)
}
)
}
label.cex.fun <- function(i, label.box.text.cex = 2) {
i <- nchar(i)
label.box.text.cex * i ^ -.25
}
}
# -------------------------
# 1: Initial Frequencies
# --------------------------
if(show.top) {
par(mar = c(0, 0, 1, 0))
plot(1, xlim = c(0, 1), ylim = c(0, 1), bty = "n", type = "n",
xlab = "", ylab = "", yaxt = "n", xaxt = "n")
if(hlines) {
segments(0, .95, 1, .95, col = panel.line.col, lwd = panel.line.lwd, lty = panel.line.lty)
}
rect(.33, .8, .67, 1.2, col = "white", border = NA)
text(x = .5, y = .95, main, cex = panel.title.cex)
text(x = .5, y = .80, paste("N = ", prettyNum(n.exemplars, big.mark = ","), "", sep = ""), cex = 1.25)
n.trueneg <- with(final.stats, cr + fa)
n.truepos <- with(final.stats, hi + mi)
text(.5, .65, paste(decision.labels[1], sep = ""), pos = 2, cex = 1.2, adj = 1)
text(.5, .65, paste(decision.labels[2], sep = ""), pos = 4, cex = 1.2, adj = 0)
#points(.9, .8, pch = 1, cex = 1.2)
#text(.9, .8, labels = paste(" = ", n.per.icon, " cases", sep = ""), pos = 4)
# Show ball examples
par(xpd = T)
add.balls.fun(x.lim = c(.35, .65),
y.lim = c(.1, .5),
n.vec = c(final.stats$fa + final.stats$cr, final.stats$hi + final.stats$mi),
pch.vec = c(noise.ball.pch, signal.ball.pch),
bg.vec = c(noise.ball.bg, signal.ball.bg),
col.vec = c(noise.ball.col, signal.ball.col),
ball.cex = ball.cex,
upper.text.adj = 2,
n.per.icon = n.per.icon
)
par(xpd = F)
# Add p.signal and p.noise levels
signal.p <- mean(x$data[[data]][[x$metadata$criterion_name]])
noise.p <- 1 - signal.p
p.rect.ylim <- c(.1, .6)
# p.signal level
text(x = .8, y = p.rect.ylim[2],
labels = paste("p(", decision.labels[2], ")", sep = ""),
pos = 3, cex = 1.2)
#Filling
rect(.775, p.rect.ylim[1],
.825, p.rect.ylim[1] + signal.p * diff(p.rect.ylim),
col = gray(.5, .25), border = NA)
# Filltop
segments(.775, p.rect.ylim[1] + signal.p * diff(p.rect.ylim),
.825, p.rect.ylim[1] + signal.p * diff(p.rect.ylim),
lwd = 1)
#Outline
rect(.775, p.rect.ylim[1],
.825, p.rect.ylim[2], lwd = 1)
if(signal.p < .0001) {signal.p.text <- "<1%"} else {
signal.p.text <- paste(round(signal.p * 100, 0), "%", sep = "")
}
text(.825, p.rect.ylim[1] + signal.p * diff(p.rect.ylim),
labels = signal.p.text,
pos = 4, cex = 1.2)
#p.noise level
text(x = .2, y = p.rect.ylim[2],
labels = paste("p(", decision.labels[1], ")", sep = ""),
pos = 3, cex = 1.2)
rect(.175, p.rect.ylim[1], .225, p.rect.ylim[1] + noise.p * diff(p.rect.ylim),
col = gray(.5, .25), border = NA)
# Filltop
segments(.175, p.rect.ylim[1] + noise.p * diff(p.rect.ylim),
.225, p.rect.ylim[1] + noise.p * diff(p.rect.ylim),
lwd = 1)
# outline
rect(.175, p.rect.ylim[1], .225, p.rect.ylim[2],
lwd = 1)
if(noise.p < .0001) {noise.p.text <- "<0.01%"} else {
noise.p.text <- paste(round(noise.p * 100, 0), "%", sep = "")
}
text(.175, p.rect.ylim[1] + noise.p * diff(p.rect.ylim),
labels = noise.p.text,
pos = 2, cex = 1.2)
}
# -------------------------
# 2. TREE
# --------------------------
if(show.middle) {
if(show.top == FALSE & show.bottom == FALSE) {
par(mar = c(3, 3, 3, 3) + .1)
} else {par(mar = c(0, 0, 0, 0))
}
# Setup plotting space
plot(1,
xlim = c(-plot.width, plot.width),
ylim = c(-plot.height, 0),
type = "n", bty = "n",
xaxt = "n", yaxt = "n",
ylab = "", xlab = ""
)
# Add frame
par(xpd = TRUE)
if(show.top | show.bottom) {
if(hlines) {
segments(-plot.width, 0, - plot.width * .3, 0, col = panel.line.col, lwd = panel.line.lwd, lty = panel.line.lty)
segments(plot.width, 0, plot.width * .3, 0, col = panel.line.col, lwd = panel.line.lwd, lty = panel.line.lty)
}
if(is.null(label.tree)) {label.tree <- paste("FFT #", tree, " (of ", x$trees$n, ")", sep = "")}
text(x = 0, y = 0,
label.tree,
cex = panel.title.cex)
}
if(show.top == FALSE & show.bottom == FALSE) {
if(is.null(main) & is.null(x$params$main)) {main <- ""}
mtext(text = main, side = 3, cex = 2)
}
par(xpd = FALSE)
# Create Noise and Signal panels
if(show.iconguide) {
# Noise Balls
points(c(- plot.width * .7, - plot.width * .5),
c(-plot.height * .125, -plot.height * .125),
pch = c(noise.ball.pch, signal.ball.pch),
bg = c(correct.bg, error.bg),
col = c(correct.border, error.border),
cex = ball.cex * 1.5
)
text(c(- plot.width * .7, - plot.width * .5),
c(-plot.height * .125, -plot.height * .125),
labels = c("Correct\nRejection", "Miss"),
pos = c(2, 4), offset = 1)
# Noise Panel
text(- plot.width * .6, -plot.height * .05,
paste("Decide ", decision.labels[1], sep = ""),
cex = 1.2, font = 3
)
# Signal panel
text(plot.width * .6, -plot.height * .05,
paste("Decide ", decision.labels[2], sep = ""),
cex = 1.2, font = 3
)
points(c(plot.width * .5, plot.width * .7),
c(-plot.height * .125, -plot.height * .125),
pch = c(noise.ball.pch, signal.ball.pch),
bg = c(error.bg, correct.bg),
col = c(error.border, correct.border),
cex = ball.cex * 1.5
)
text(c(plot.width * .5, plot.width * .7),
c(-plot.height * .125, -plot.height * .125),
labels = c("False\nAlarm", "Hit"),
pos = c(2, 4), offset = 1)
}
# Set initial subplot center
subplot.center <- c(0, -4)
# Loop over levels
for(level.i in 1:min(c(n.levels, 6))) {
current.cue <- cue.labels[level.i]
# Get stats for current level
{
hi.i <- level.stats$hi.m[level.i]
mi.i <- level.stats$mi.m[level.i]
fa.i <- level.stats$fa.m[level.i]
cr.i <- level.stats$cr.m[level.i]
}
# If level.i == 1, draw top textbox
{
if(level.i == 1) {
rect(subplot.center[1] - label.box.width / 2,
subplot.center[2] + 2 - label.box.height / 2,
subplot.center[1] + label.box.width / 2,
subplot.center[2] + 2 + label.box.height / 2,
col = "white",
border = "black"
)
points(x = subplot.center[1],
y = subplot.center[2] + 2,
cex = decision.node.cex,
pch = decision.node.pch
)
text(x = subplot.center[1],
y = subplot.center[2] + 2,
labels = current.cue,
cex = label.box.text.cex#label.cex.fun(current.cue, label.box.text.cex = label.box.text.cex)
)
}
}
# -----------------------
# Left (Noise) Classification / New Level
# -----------------------
{
# Exit node on left
if(level.stats$exit[level.i] %in% c(0, .5) | paste(level.stats$exit[level.i]) %in% c("0", ".5")) {
segments(subplot.center[1],
subplot.center[2] + 1,
subplot.center[1] - 2,
subplot.center[2] - 2,
lty = segment.lty,
lwd = segment.lwd
)
arrows(x0 = subplot.center[1] - 2,
y0 = subplot.center[2] - 2,
x1 = subplot.center[1] - 2 - arrow.length,
y1 = subplot.center[2] - 2,
lty = arrow.lty,
lwd = arrow.lwd,
col = arrow.col,
length = arrow.head.length
)
# Decision text
if(decision.cex > 0) {
text(x = subplot.center[1] - 2 - arrow.length * .7,
y = subplot.center[2] - 2.2,
labels = decision.labels[1],
pos = 1, font = 3, cex = decision.cex)
}
if(ball.loc == "fixed") {
ball.x.lim <- c(-max(ball.box.fixed.x.shift), -min(ball.box.fixed.x.shift))
ball.y.lim <- c(subplot.center[2] + ball.box.vert.shift - ball.box.height / 2,
subplot.center[2] + ball.box.vert.shift + ball.box.height / 2)
}
if(ball.loc == "variable") {
ball.x.lim <- c(subplot.center[1] - ball.box.horiz.shift - ball.box.width / 2,
subplot.center[1] - ball.box.horiz.shift + ball.box.width / 2)
ball.y.lim <- c(subplot.center[2] + ball.box.vert.shift - ball.box.height / 2,
subplot.center[2] + ball.box.vert.shift + ball.box.height / 2)
}
if(max(c(cr.i, mi.i), na.rm = TRUE) > 0 & show.icons == TRUE) {
add.balls.fun(x.lim = ball.x.lim,
y.lim = ball.y.lim,
n.vec = c(cr.i, mi.i),
pch.vec = c(noise.ball.pch, signal.ball.pch),
# bg.vec = c(noise.ball.bg, signal.ball.bg),
bg.vec = c(correct.bg, error.bg),
col.vec = c(correct.border, error.border),
freq.text = TRUE,
n.per.icon = n.per.icon,
ball.cex = ball.cex
)
}
# level break label
pos.direction.symbol <- c("<=", "<", "=", "!=", ">", ">=")[which(level.stats$direction[level.i] == c(">", ">=", "!=", "=", "<=", "<"))]
neg.direction.symbol <- c("<=", "<", "=", "!=", ">", ">=")[which(level.stats$direction[level.i] == c("<=", "<", "=", "!=", ">", ">="))]
text.outline(x = subplot.center[1] - 1,
y = subplot.center[2],
labels = paste(pos.direction.symbol, " ", level.stats$threshold[level.i], sep = ""),
pos = 2, cex = break.label.cex, r = .1
)
points(x = subplot.center[1] - 2,
y = subplot.center[2] - 2,
pch = exit.node.pch,
cex = exit.node.cex,
bg = exit.node.bg
)
text(x = subplot.center[1] - 2,
y = subplot.center[2] - 2,
labels = substr(decision.labels[1], 1, 1)
)
}
# New level on left
if(level.stats$exit[level.i] %in% c(1) | paste(level.stats$exit[level.i]) %in% c("1")) {
segments(subplot.center[1],
subplot.center[2] + 1,
subplot.center[1] - 2,
subplot.center[2] - 2,
lty = segment.lty,
lwd = segment.lwd
)
rect(subplot.center[1] - 2 - label.box.width / 2,
subplot.center[2] - 2 - label.box.height / 2,
subplot.center[1] - 2 + label.box.width / 2,
subplot.center[2] - 2 + label.box.height / 2,
col = "white",
border = "black"
)
if(level.i < 6) {text(x = subplot.center[1] - 2,
y = subplot.center[2] - 2,
labels = cue.labels[level.i + 1],
cex = label.box.text.cex
)} else {
text(x = subplot.center[1] - 2,
y = subplot.center[2] - 2,
labels = paste0("+ ", n.levels - 6, " More"),
cex = label.box.text.cex,
font = 3
)
}
}
}
# -----------------------
# Right (Signal) Classification / New Level
# -----------------------
{
# Exit node on right
if(level.stats$exit[level.i] %in% c(1, .5) | paste(level.stats$exit[level.i]) %in% c("1", ".5")) {
segments(subplot.center[1],
subplot.center[2] + 1,
subplot.center[1] + 2,
subplot.center[2] - 2,
lty = segment.lty,
lwd = segment.lwd
)
arrows(x0 = subplot.center[1] + 2,
y0 = subplot.center[2] - 2,
x1 = subplot.center[1] + 2 + arrow.length,
y1 = subplot.center[2] - 2,
lty = arrow.lty,
lwd = arrow.lwd,
col = arrow.col,
length = arrow.head.length
)
# Decision text
if(decision.cex > 0) {
text(x = subplot.center[1] + 2 + arrow.length * .7,
y = subplot.center[2] - 2.2,
labels = decision.labels[2],
pos = 1, font = 3, cex = decision.cex)
}
if(ball.loc == "fixed") {
ball.x.lim <- c(min(ball.box.fixed.x.shift), max(ball.box.fixed.x.shift))
ball.y.lim <- c(subplot.center[2] + ball.box.vert.shift - ball.box.height / 2,
subplot.center[2] + ball.box.vert.shift + ball.box.height / 2)
}
if(ball.loc == "variable") {
ball.x.lim <- c(subplot.center[1] + ball.box.horiz.shift - ball.box.width / 2,
subplot.center[1] + ball.box.horiz.shift + ball.box.width / 2)
ball.y.lim <- c(subplot.center[2] + ball.box.vert.shift - ball.box.height / 2,
subplot.center[2] + ball.box.vert.shift + ball.box.height / 2)
}
if(max(c(fa.i, hi.i), na.rm = TRUE) > 0 & show.icons == TRUE) {
add.balls.fun(x.lim = ball.x.lim,
y.lim = ball.y.lim,
n.vec = c(fa.i, hi.i),
pch.vec = c(noise.ball.pch, signal.ball.pch),
# bg.vec = c(noise.ball.bg, signal.ball.bg),
bg.vec = c(error.bg, correct.bg),
col.vec = c(error.border, correct.border),
freq.text = TRUE,
n.per.icon = n.per.icon,
ball.cex = ball.cex
)
}
# level break label
dir.symbols <- c("<=", "<", "=", "!=", ">", ">=")
pos.direction.symbol <- dir.symbols[which(level.stats$direction[level.i] == c("<=", "<", "=", "!=", ">", ">="))]
neg.direction.symbol <- dir.symbols[which(level.stats$direction[level.i] == rev(c("<=", "<", "=", "!=", ">", ">=")))]
text.outline(subplot.center[1] + 1,
subplot.center[2],
labels = paste(pos.direction.symbol, " ", level.stats$threshold[level.i], sep = ""),
pos = 4, cex = break.label.cex, r = .1
)
points(x = subplot.center[1] + 2,
y = subplot.center[2] - 2,
pch = exit.node.pch,
cex = exit.node.cex,
bg = exit.node.bg
)
text(x = subplot.center[1] + 2,
y = subplot.center[2] - 2,
labels = substr(decision.labels[2], 1, 1)
)
}
# New level on right
if(level.stats$exit[level.i] %in% 0 | paste(level.stats$exit[level.i]) %in% c("0")) {
segments(subplot.center[1],
subplot.center[2] + 1,
subplot.center[1] + 2,
subplot.center[2] - 2,
lty = segment.lty,
lwd = segment.lwd
)
if(level.i < 6) {
rect(subplot.center[1] + 2 - label.box.width / 2,
subplot.center[2] - 2 - label.box.height / 2,
subplot.center[1] + 2 + label.box.width / 2,
subplot.center[2] - 2 + label.box.height / 2,
col = "white",
border = "black")
text(x = subplot.center[1] + 2,
y = subplot.center[2] - 2,
labels = cue.labels[level.i + 1],
cex = label.box.text.cex)
} else {
rect(subplot.center[1] + 2 - label.box.width / 2,
subplot.center[2] - 2 - label.box.height / 2,
subplot.center[1] + 2 + label.box.width / 2,
subplot.center[2] - 2 + label.box.height / 2,
col = "white",
border = "black", lty = 2)
text(x = subplot.center[1] + 2,
y = subplot.center[2] - 2,
labels = paste0("+ ", n.levels - 6, " More"),
cex = label.box.text.cex,
font = 3
)
}
}
}
# -----------------------
# Update plot center
# -----------------------
{
if(level.stats$exit[level.i] == 0) {
subplot.center <- c(subplot.center[1] + 2,
subplot.center[2] - 4)
}
if(level.stats$exit[level.i] == 1) {
subplot.center <- c(subplot.center[1] - 2,
subplot.center[2] - 4)
}
}
}
}
# -----------------------
# 3. CUMULATIVE PERFORMANCE
# -----------------------
if(show.bottom == TRUE) {
{
# OBTAIN FINAL STATISTICS
fft.sens.vec <- tree.stats$sens
fft.spec.vec <- tree.stats$spec
# General plotting space
{
# PLOTTING PARAMETERS
header.y.loc <- 1.0
subheader.y.loc <- .925
header.cex <- 1.1
subheader.cex <- .9
par(mar = c(0, 0, 2, 0))
plot(1, xlim = c(0, 1), ylim = c(0, 1),
bty = "n", type = "n",
xlab = "", ylab = "",
yaxt = "n", xaxt = "n")
par(xpd = T)
if(hlines) {
segments(0, 1.1, 1, 1.1, col = panel.line.col, lwd = panel.line.lwd, lty = panel.line.lty)
rect(.25, 1, .75, 1.2, col = "white", border = NA)
}
if(is.null(label.performance)) {
if(data == "train") {label.performance <- "Accuracy (Training)"}
if(data == "test") {label.performance <- "Accuracy (Testing)"}
}
text(.5, 1.1, label.performance, cex = panel.title.cex)
par(xpd = FALSE)
pretty.dec <- function(x) {return(paste(round(x, 2) * 100, sep = ""))}
level.max.height <- .65
level.width <- .05
level.center.y <- .45
#level.bottom <- .1
level.bottom <- level.center.y - level.max.height / 2
level.top <- level.center.y + level.max.height / 2
lloc <- data.frame(
element = c("classtable", "mcu", "pci", "sens", "spec", "acc", "bacc", "roc"),
long.name = c("Classification Table", "mcu", "pci", "sens", "spec", "acc", "bacc", "ROC"),
center.x = c(.18, seq(.35, .65, length.out = 6), .85),
center.y = rep(level.center.y, 8),
width = c(.2, rep(level.width, 6), .2),
height = c(.65, rep(level.max.height, 6), .65),
value = c(NA,
abs(final.stats$mcu - 5) / (abs(1 - 5)),
final.stats$pci, final.stats$sens, final.stats$spec, with(final.stats, (cr + hi) / n), final.stats$bacc, NA),
value.name = c(NA, round(final.stats$mcu, 1), pretty.dec(final.stats$pci), pretty.dec(final.stats$sens), pretty.dec(final.stats$spec), pretty.dec(final.stats$acc),
pretty.dec(final.stats$bacc), NA
)
)
}
# Classification table
if(show.confusion) {
final.classtable.x.loc <- c(lloc$center.x[lloc$element == "classtable"] - lloc$width[lloc$element == "classtable"] / 2, lloc$center.x[lloc$element == "classtable"] + lloc$width[lloc$element == "classtable"] / 2)
final.classtable.y.loc <- c(lloc$center.y[lloc$element == "classtable"] - lloc$height[lloc$element == "classtable"] / 2, lloc$center.y[lloc$element == "classtable"] + lloc$height[lloc$element == "classtable"] / 2)
rect(final.classtable.x.loc[1], final.classtable.y.loc[1],
final.classtable.x.loc[2], final.classtable.y.loc[2],
lwd = classtable.lwd
)
segments(mean(final.classtable.x.loc), final.classtable.y.loc[1], mean(final.classtable.x.loc), final.classtable.y.loc[2], col = gray(0), lwd = classtable.lwd)
segments(final.classtable.x.loc[1], mean(final.classtable.y.loc), final.classtable.x.loc[2], mean(final.classtable.y.loc), col = gray(0), lwd = classtable.lwd)
# Column titles
text(x = mean(mean(final.classtable.x.loc)),
y = header.y.loc,
"Truth", pos = 1, cex = header.cex)
text(x = final.classtable.x.loc[1] + .25 * diff(final.classtable.x.loc),
y = subheader.y.loc, pos = 1, cex = subheader.cex,
decision.labels[2])
text(x = final.classtable.x.loc[1] + .75 * diff(final.classtable.x.loc),
y = subheader.y.loc, pos = 1, cex = subheader.cex,
decision.labels[1])
# Row titles
text(x = final.classtable.x.loc[1] - .01,
y = final.classtable.y.loc[1] + .75 * diff(final.classtable.y.loc), cex = subheader.cex,
decision.labels[2], adj = 1)
text(x = final.classtable.x.loc[1] - .01,
y = final.classtable.y.loc[1] + .25 * diff(final.classtable.y.loc), cex = subheader.cex,
decision.labels[1], adj = 1)
text(x = final.classtable.x.loc[1] - .065,
y = mean(final.classtable.y.loc), cex = header.cex,
"Decision")
# text(x = final.classtable.x.loc[1] - .05,
# y = mean(final.classtable.y.loc), cex = header.cex,
# "Decision", srt = 90, pos = 3)
# Add final frequencies
text(final.classtable.x.loc[1] + .75 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .25 * diff(final.classtable.y.loc),
prettyNum(final.stats$cr, big.mark = ","), cex = 1.5)
text(final.classtable.x.loc[1] + .25 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .25 * diff(final.classtable.y.loc),
prettyNum(final.stats$mi, big.mark = ","), cex = 1.5)
text(final.classtable.x.loc[1] + .75 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .75 * diff(final.classtable.y.loc),
prettyNum(final.stats$fa, big.mark = ","), cex = 1.5)
text(final.classtable.x.loc[1] + .25 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .75 * diff(final.classtable.y.loc),
prettyNum(final.stats$hi, big.mark = ","), cex = 1.5)
# Add symbols
points(final.classtable.x.loc[1] + .55 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .05 * diff(final.classtable.y.loc),
pch = noise.ball.pch, bg = correct.bg, col = correct.border, cex = ball.cex)
points(final.classtable.x.loc[1] + .05 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .55 * diff(final.classtable.y.loc),
pch = signal.ball.pch, bg = correct.bg, cex = ball.cex, col = correct.border)
points(final.classtable.x.loc[1] + .55 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .55 * diff(final.classtable.y.loc),
pch = noise.ball.pch, bg = error.bg, col = error.border, cex = ball.cex)
points(final.classtable.x.loc[1] + .05 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .05 * diff(final.classtable.y.loc),
pch = signal.ball.pch, bg = error.bg, col = error.border, cex = ball.cex)
# Labels
text(final.classtable.x.loc[1] + .62 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .07 * diff(final.classtable.y.loc),
"cr", cex = 1, font = 3, adj = 0)
text(final.classtable.x.loc[1] + .12 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .07 * diff(final.classtable.y.loc),
"mi", cex = 1, font = 3, adj = 0)
text(final.classtable.x.loc[1] + .62 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .57 * diff(final.classtable.y.loc),
"fa", cex = 1, font = 3, adj = 0)
text(final.classtable.x.loc[1] + .12 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .57 * diff(final.classtable.y.loc),
"hi", cex = 1, font = 3, adj = 0)
}
# Levels
if(show.levels) {
if(level.type %in% c("line", "bar")) {
# Color function (taken from colorRamp2 function in circlize package)
col.fun <- circlize::colorRamp2(c(0, .75, 1), c("red", "yellow", "green"), transparency = .1)
add.level.fun <- function(name,
sub = "",
max.val = 1,
min.val = 0,
ok.val = .5,
bottom.text = "",
level.type = "line") {
rect.center.x <- lloc$center.x[lloc$element == name]
rect.center.y <- lloc$center.y[lloc$element == name]
rect.height <- lloc$height[lloc$element == name]
rect.width <- lloc$width[lloc$element == name]
rect.bottom.y <- rect.center.y - rect.height / 2
rect.top.y <- rect.center.y + rect.height / 2
rect.left.x <- rect.center.x - rect.width / 2
rect.right.x <- rect.center.x + rect.width / 2
long.name <- lloc$long.name[lloc$element == name]
value <- lloc$value[lloc$element == name]
value.name <- lloc$value.name[lloc$element == name]
#
# level.col.fun <- circlize::colorRamp2(c(min.val, ok.val, max.val),
# colors = c("firebrick2", "yellow", "green4"),
# transparency = .1)
text(x = rect.center.x,
y = header.y.loc,
labels = long.name,
pos = 1, cex = header.cex)
#
# text.outline(x = rect.center.x,
# y = header.y.loc,
# labels = long.name,
# pos = 1, cex = header.cex, r = .02
# )
value.height <- rect.bottom.y + min(c(1, ((value - min.val) / (max.val - min.val)))) * rect.height
# Add filling
value.s <- min(value / max.val, 1)
delta <- 1
gamma <- .5
value.col.scale <- delta * value.s ^ gamma / (delta * value.s ^ gamma + (1 - value.s) ^ gamma)
# value.col <- gray(1 - value.col.scale * .5)
value.col <- gray(1, .25)
#plot(seq(0, 1, .01), delta * seq(0, 1, .01) ^ gamma / (delta * seq(0, 1, .01) ^ gamma + (1 - seq(0, 1, .01)) ^ gamma))
if(level.type == "bar") {
rect(rect.left.x,
rect.bottom.y,
rect.right.x,
value.height,
#col = level.col.fun(value.s),
col = value.col,
# col = spec.level.fun(lloc$value[lloc$element == name]),
border = "black"
)
text.outline(x = rect.center.x,
y = value.height,
labels = lloc$value.name[lloc$element == name],
cex = 1.5, r = .008, pos = 3
)
# Add level border
# rect(rect.left.x,
# rect.bottom.y,
# rect.right.x,
# rect.top.y,
# border = gray(.5, .5))
}
if(level.type == "line") {
# Stem
segments(rect.center.x,
rect.bottom.y,
rect.center.x,
value.height,
lty = 3)
# Horizontal platform
platform.width <- .02
segments(rect.center.x - platform.width,
value.height,
rect.center.x + platform.width,
value.height)
# Text label
text.outline(x = rect.center.x,
y = value.height,
labels = lloc$value.name[lloc$element == name],
cex = 1.5, r = 0, pos = 3
)
# points(rect.center.x,
# value.height,
# cex = 5.5,
# pch = 21,
# bg = "white",
# col = "black", lwd = .5)
}
# Add subtext
text(x = rect.center.x,
y = rect.center.y - .05,
labels = sub,
cex = .8,
font = 1,
pos = 1)
# Add bottom text
text(x = rect.center.x,
y = rect.bottom.y,
labels = bottom.text,
pos = 1)
}
paste(final.stats$cr, "/", 1, collapse = "")
#Add 100% reference line
# segments(x0 = lloc$center.x[lloc$element == "mcu"] - lloc$width[lloc$element == "mcu"] * .8,
# y0 = level.top,
# x1 = lloc$center.x[lloc$element == "bacc"] + lloc$width[lloc$element == "bacc"] * .8,
# y1 = level.top,
# lty = 3, lwd = .75)
add.level.fun("mcu",
ok.val = .75,
max.val = 1,
min.val = 0,
level.type = level.type) #, sub = paste(c(final.stats$cr, "/", final.stats$cr + final.stats$fa), collapse = ""))
add.level.fun("pci", ok.val = .75, level.type = level.type) #, sub = paste(c(final.stats$cr, "/", final.stats$cr + final.stats$fa), collapse = ""))
# text(lloc$center.x[lloc$element == "pci"],
# lloc$center.y[lloc$element == "pci"],
# labels = paste0("mcu\n", round(mcu, 2)))
add.level.fun("spec", ok.val = .75, level.type = level.type) #, sub = paste(c(final.stats$cr, "/", final.stats$cr + final.stats$fa), collapse = ""))
add.level.fun("sens", ok.val = .75, level.type = level.type) #, sub = paste(c(final.stats$hi, "/", final.stats$hi + final.stats$mi), collapse = ""))
# Min acc
min.acc <- max(crit.br, 1 - crit.br)
add.level.fun("acc", min.val = 0, ok.val = .5, level.type = level.type) #, sub = paste(c(final.stats$hi + final.stats$cr, "/", final.stats$n), collapse = ""))
# Add baseline to acc level
segments(x0 = lloc$center.x[lloc$element == "acc"] - lloc$width[lloc$element == "acc"] / 2,
y0 = (lloc$center.y[lloc$element == "acc"] - lloc$height[lloc$element == "acc"] / 2) + lloc$height[lloc$element == "acc"] * min.acc,
x1 = lloc$center.x[lloc$element == "acc"] + lloc$width[lloc$element == "acc"] / 2,
y1 = (lloc$center.y[lloc$element == "acc"] - lloc$height[lloc$element == "acc"] / 2) + lloc$height[lloc$element == "acc"] * min.acc,
lty = 3)
text(x = lloc$center.x[lloc$element == "acc"],
y =(lloc$center.y[lloc$element == "acc"] - lloc$height[lloc$element == "acc"] / 2) + lloc$height[lloc$element == "acc"] * min.acc,
labels = "BL", pos = 1)
# paste("BL = ", pretty.dec(min.acc), sep = ""), pos = 1)
add.level.fun("bacc", min.val = 0, max.val = 1, ok.val = .5, level.type = level.type)
# baseline
# segments(x0 = mean(lloc$center.x[2]),
# y0 = lloc$center.y[1] - lloc$height[1] / 2,
# x1 = mean(lloc$center.x[7]),
# y1 = lloc$center.y[1] - lloc$height[1] / 2, lend = 1,
# lwd = .5,
# col = gray(0))
}
}
# MiniROC
if(show.roc) {
text(lloc$center.x[lloc$element == "roc"], header.y.loc, "ROC", pos = 1, cex = header.cex)
# text(final.roc.center[1], subheader.y.loc, paste("AUC =", round(final.auc, 2)), pos = 1)
final.roc.x.loc <- c(lloc$center.x[lloc$element == "roc"] - lloc$width[lloc$element == "roc"] / 2,lloc$center.x[lloc$element == "roc"] + lloc$width[lloc$element == "roc"] / 2)
final.roc.y.loc <- c(lloc$center.y[lloc$element == "roc"] - lloc$height[lloc$element == "roc"] / 2,lloc$center.y[lloc$element == "roc"] + lloc$height[lloc$element == "roc"] / 2)
# Plot bg
#
# rect(final.roc.x.loc[1],
# final.roc.y.loc[1],
# final.roc.x.loc[2],
# final.roc.y.loc[2],
# col = gray(1), lwd = .5)
# Gridlines
# # Horizontal
# segments(x0 = rep(final.roc.x.loc[1], 9),
# y0 = seq(final.roc.y.loc[1], final.roc.y.loc[2], length.out = 5)[2:10],
# x1 = rep(final.roc.x.loc[2], 9),
# y1 = seq(final.roc.y.loc[1], final.roc.y.loc[2], length.out = 5)[2:10],
# lty = 1, col = gray(.8), lwd = c(.5), lend = 3
# )
#
# # Vertical
# segments(y0 = rep(final.roc.y.loc[1], 9),
# x0 = seq(final.roc.x.loc[1], final.roc.x.loc[2], length.out = 5)[2:10],
# y1 = rep(final.roc.y.loc[2], 9),
# x1 = seq(final.roc.x.loc[1], final.roc.x.loc[2], length.out = 5)[2:10],
# lty = 1, col = gray(.8), lwd = c(.5), lend = 3
# )
# Plot border
rect(final.roc.x.loc[1],
final.roc.y.loc[1],
final.roc.x.loc[2],
final.roc.y.loc[2],
border = roc.border.col,
lwd = roc.lwd)
# Axis labels
text(c(final.roc.x.loc[1], final.roc.x.loc[2]),
c(final.roc.y.loc[1], final.roc.y.loc[1]) - .05,
labels = c(0, 1))
text(c(final.roc.x.loc[1], final.roc.x.loc[1], final.roc.x.loc[1]) - .02,
c(final.roc.y.loc[1], mean(final.roc.y.loc[1:2]), final.roc.y.loc[2]),
labels = c(0,.5, 1))
text(mean(final.roc.x.loc), final.roc.y.loc[1] - .08, "1 - Specificity (FAR)")
text(final.roc.x.loc[1] - .04, mean(final.roc.y.loc), "Sensitivity (HR)", srt = 90)
# Diagonal
segments(final.roc.x.loc[1],
final.roc.y.loc[1],
final.roc.x.loc[2],
final.roc.y.loc[2],
lty = 3)
label.loc <- c(.1, .3, .5, .7, .9)
## COMPETITIVE ALGORITHMS
if(comp == TRUE) {
# CART
if(is.null(x$comp$cart$results) == FALSE) {
cart.spec <- x$comp$cart$results[[data]]$spec
cart.sens <- x$comp$cart$results[[data]]$sens
points(final.roc.x.loc[1] + (1 - cart.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + cart.sens * lloc$height[lloc$element == "roc"],
pch = 21, cex = 1.75,
col = yarrr::transparent("red", .1),
bg = yarrr::transparent("red", .9), lwd = 1)
points(final.roc.x.loc[1] + (1 - cart.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + cart.sens * lloc$height[lloc$element == "roc"],
pch = "C", cex = .7, col = gray(.2), lwd = 1)
par("xpd" = FALSE)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[4] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5,
col = yarrr::transparent("red", .1),
bg = yarrr::transparent("red", .9))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[4] * lloc$height[lloc$element == "roc"],
pch = "C", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[4] * lloc$height[lloc$element == "roc"],
labels = " CART", adj = 0, cex = .9)
par("xpd" = TRUE)
}
## LR
if(is.null(x$comp$lr$results) == FALSE) {
lr.spec <- x$comp$lr$results[[data]]$spec
lr.sens <- x$comp$lr$results[[data]]$sens
points(final.roc.x.loc[1] + (1 - lr.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + lr.sens * lloc$height[lloc$element == "roc"],
pch = 21, cex = 1.75,
col = yarrr::transparent("blue", .1),
bg = yarrr::transparent("blue", .9))
points(final.roc.x.loc[1] + (1 - lr.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + lr.sens * lloc$height[lloc$element == "roc"],
pch = "L", cex = .7, col = gray(.2))
par("xpd" = F)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[3] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5,
col = yarrr::transparent("blue", .1),
bg = yarrr::transparent("blue", .9))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[3] * lloc$height[lloc$element == "roc"],
pch = "L", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[3] * lloc$height[lloc$element == "roc"],
labels = " LR", adj = 0, cex = .9)
par("xpd" = T)
}
## rf
if(is.null(x$comp$rf$results) == FALSE) {
rf.spec <- x$comp$rf$results[[data]]$spec
rf.sens <- x$comp$rf$results[[data]]$sens
points(final.roc.x.loc[1] + (1 - rf.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + rf.sens * lloc$height[lloc$element == "roc"],
pch = 21, cex = 1.75,
col = yarrr::transparent("purple", .1),
bg = yarrr::transparent("purple", .9), lwd = 1)
points(final.roc.x.loc[1] + (1 - rf.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + rf.sens * lloc$height[lloc$element == "roc"],
pch = "C", cex = .7, col = gray(.2), lwd = 1)
par("xpd" = FALSE)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[2] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5,
col = yarrr::transparent("purple", .1),
bg = yarrr::transparent("purple", .9))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[2] * lloc$height[lloc$element == "roc"],
pch = "R", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[2] * lloc$height[lloc$element == "roc"],
labels = " RF", adj = 0, cex = .9)
par("xpd" = TRUE)
}
## svm
## svm
if(is.null(x$comp$svm$results) == FALSE) {
svm.spec <- x$comp$svm$results[[data]]$spec
svm.sens <- x$comp$svm$results[[data]]$sens
points(final.roc.x.loc[1] + (1 - svm.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + svm.sens * lloc$height[lloc$element == "roc"],
pch = 21, cex = 1.75,
col = yarrr::transparent("orange", .1),
bg = yarrr::transparent("orange", .9), lwd = 1)
points(final.roc.x.loc[1] + (1 - svm.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + svm.sens * lloc$height[lloc$element == "roc"],
pch = "C", cex = .7, col = gray(.2), lwd = 1)
par("xpd" = FALSE)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[1] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5,
col = yarrr::transparent("orange", .1),
bg = yarrr::transparent("orange", .9))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[1] * lloc$height[lloc$element == "roc"],
pch = "S", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[1] * lloc$height[lloc$element == "roc"],
labels = " SVM", adj = 0, cex = .9)
par("xpd" = TRUE)
}
}
## FFT
{
roc.order <- order(fft.spec.vec, decreasing = TRUE)
# roc.order <- 1:x$trees$n
fft.sens.vec.ord <- fft.sens.vec[roc.order]
fft.spec.vec.ord <- fft.spec.vec[roc.order]
# Add segments and points for all trees but tree
if(length(roc.order) > 1) {
segments(final.roc.x.loc[1] + c(0, 1 - fft.spec.vec.ord) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + c(0, fft.sens.vec.ord) * lloc$height[lloc$element == "roc"],
final.roc.x.loc[1] + c(1 - fft.spec.vec.ord, 1) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + c(fft.sens.vec.ord, 1) * lloc$height[lloc$element == "roc"],
lwd = 1,
col = gray(0))
points(final.roc.x.loc[1] + (1 - fft.spec.vec.ord[-(which(roc.order == tree))]) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + fft.sens.vec.ord[-(which(roc.order == tree))] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5, col = yarrr::transparent("green", .3),
bg = yarrr::transparent("white", .1))
text(final.roc.x.loc[1] + (1 - fft.spec.vec.ord[-(which(roc.order == tree))]) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + fft.sens.vec.ord[-(which(roc.order == tree))] * lloc$height[lloc$element == "roc"],
labels = roc.order[which(roc.order != tree)], cex = 1, col = gray(.2))
}
# Add large point for plotted tree
points(final.roc.x.loc[1] + (1 - fft.spec.vec[tree]) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + fft.sens.vec[tree] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 3, col = gray(1), #col = yarrr::transparent("green", .3),
bg = yarrr::transparent("green", .3), lwd = 1)
text(final.roc.x.loc[1] + (1 - fft.spec.vec[tree]) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + fft.sens.vec[tree] * lloc$height[lloc$element == "roc"],
labels = tree, cex = 1.25, col = gray(.2), font = 2)
# Labels
if(comp == TRUE & any(is.null(x$comp$lr$results) == FALSE,
is.null(x$comp$cart$results) == FALSE,
is.null(x$comp$svm$results) == FALSE,
is.null(x$comp$rf$results) == FALSE
)) {
par("xpd" = FALSE)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[5] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5, col = yarrr::transparent("green", .3),
bg = yarrr::transparent("green", .7))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[5] * lloc$height[lloc$element == "roc"],
pch = "#", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[5] * lloc$height[lloc$element == "roc"],
labels = " FFT", adj = 0, cex = .9)
par("xpd" = TRUE)
}
}
}
}
}
# Reset plotting space
par(mfrow = c(1, 1))
par(mar = c(5, 4, 4, 1) + .1)
}
}
| /R/plotFFTrees_function.R | no_license | guhjy/FFTrees | R | false | false | 57,366 | r | #' Plots an FFTrees object.
#'
#' @description Plots an FFTrees object created by the FFTrees() function.
#' @param x A FFTrees object created from \code{"FFTrees()"}
#' @param data One of two strings 'train' or 'test'. In this case, the corresponding dataset in the x object will be used.
#' @param what string. What should be plotted? \code{'tree'} (the default) shows one tree (specified by \code{'tree'}). \code{'cues'} shows the marginal accuracy of cues in an ROC space, \code{"roc"} shows an roc curve of the tree(s)
#' @param tree integer. An integer indicating which tree to plot (only valid when the tree argument is non-empty). To plot the best training (or test) tree with respect to the \code{goal} specified during FFT construction, use "best.train" or "best.test"
#' @param main character. The main plot label.
#' @param hlines logical. Should horizontal panel separation lines be shown?
#' @param cue.labels character. An optional string of labels for the cues / nodes.
#' @param decision.labels character. A string vector of length 2 indicating the content-specific name for noise and signal cases.
#' @param cue.cex numeric. The size of the cue labels.
#' @param threshold.cex numeric. The size of the threshold labels.
#' @param decision.cex numeric. The size of the decision labels.
#' @param comp logical. Should the performance of competitive algorithms (e.g.; logistic regression, random forests etc.) be shown in the ROC plot (if available?)
#' @param stats logical. Should statistical information be plotted? If \code{FALSE}, then only the tree (without any reference to statistics) will be plotted.
#' @param show.header,show.tree,show.confusion,show.levels,show.roc,show.icons,show.iconguide logical. Logical arguments indicating which specific elements of the plot to show.
#' @param label.tree,label.performance string. Optional arguments to define lables for the tree and performance section(s).
#' @param n.per.icon Number of cases per icon
#' @param which.tree deprecated argument, only for backwards compatibility, use \code{"tree"} instead.
#' @param level.type string. How should bottom levels be drawn? Can be \code{"bar"} or \code{"line"}
#' @param decision.names deprecated arguments.
#' @param ... Currently ignored.
#' @importFrom stats anova predict formula model.frame
#' @importFrom graphics text points abline legend mtext segments rect arrows axis par layout plot
#' @importFrom grDevices gray col2rgb rgb
#' @importFrom yarrr transparent piratepal
#' @export
#' @examples
#'
#' # Create FFTrees of the heart disease data
#' heart.fft <- FFTrees(formula = diagnosis ~.,
#' data = heartdisease)
#'
#' # Visualise the tree
#' plot(heart.fft,
#' main = "Heart Disease Diagnosis",
#' decision.labels = c("Absent", "Present"))
#'
#'
#' # See the vignette for more details
#' vignette("FFTrees_plot", package = "FFTrees")
#'
#'
#'
plot.FFTrees <- function(
x = NULL,
data = "train",
what = 'tree',
tree = "best.train",
main = NULL,
hlines = TRUE,
cue.labels = NULL,
decision.labels = NULL,
cue.cex = NULL,
threshold.cex = NULL,
decision.cex = 1,
comp = TRUE,
stats = TRUE,
show.header = NULL,
show.tree = NULL,
show.confusion = NULL,
show.levels = NULL,
show.roc = NULL,
show.icons = NULL,
show.iconguide = NULL,
label.tree = NULL,
label.performance = NULL,
n.per.icon = NULL,
which.tree = NULL,
level.type = "bar",
decision.names = NULL,
...
) {
#
# # #
# data = "train"
# what = 'tree'
# tree = "best.train"
# main = NULL
# hlines = TRUE
# cue.labels = NULL
# decision.labels = NULL
# cue.cex = NULL
# threshold.cex = NULL
# decision.cex = 1
# comp = TRUE
# stats = TRUE
# show.header = NULL
# show.tree = NULL
# show.confusion = NULL
# show.levels = NULL
# show.roc = NULL
# show.icons = NULL
# show.iconguide = NULL
# label.tree = NULL
# label.performance = NULL
# n.per.icon = NULL
# which.tree = NULL
# level.type = "bar"
# decision.names = NULL
# Check for invalid or missing arguments
# Input validation
{
if(what %in% c("cues", "tree", "roc") == FALSE) {
stop("what must be either 'cues', 'tree', or 'roc'")
}
if(is.null(decision.names) == FALSE) {
message("decision.names is deprecated, use decision.lables instead")
decision.labels <- decision.names
}
}
# If what == cues, then send inputs to showcues()
if(what == 'cues') {showcues(x = x, data = data, main = main)}
if(what != 'cues') {
# Determine layout
{
if(what == "tree") {
if(stats == TRUE) {
if(is.null(show.header)) {show.header <- TRUE}
if(is.null(show.tree)) {show.tree <- TRUE}
if(is.null(show.confusion)) {show.confusion <- TRUE}
if(is.null(show.levels)) {show.levels <- TRUE}
if(is.null(show.roc)) {show.roc <- TRUE}
if(is.null(show.icons)) {show.icons <- TRUE}
if(is.null(show.iconguide)) {show.iconguide <- TRUE}
}
if(stats == FALSE) {
if(is.null(show.header)) {show.header <- FALSE}
if(is.null(show.tree)) {show.tree <- TRUE}
if(is.null(show.confusion)) {show.confusion <- FALSE}
if(is.null(show.levels)) {show.levels <- FALSE}
if(is.null(show.roc)) {show.roc <- FALSE}
if(is.null(show.icons)) {show.icons <- FALSE}
if(is.null(show.iconguide)) {show.iconguide <- FALSE}
}
}
if(what == "roc") {
show.header <- FALSE
show.tree <- FALSE
show.confusion <- FALSE
show.levels <- FALSE
show.roc <- TRUE
show.icons <- FALSE
show.top <- FALSE
}
# Top, middle, bottom
if(show.header & show.tree & (show.confusion | show.levels | show.roc)) {
show.top <- TRUE
show.middle <- TRUE
show.bottom <- TRUE
layout(matrix(1:3, nrow = 3, ncol = 1),
widths = c(6),
heights = c(1.2, 3, 1.8))
}
# Top and middle
if(show.header & show.tree & (show.confusion == FALSE & show.levels == FALSE & show.roc == FALSE)) {
show.top <- TRUE
show.middle <- TRUE
show.bottom <- FALSE
layout(matrix(1:2, nrow = 2, ncol = 1),
widths = c(6),
heights = c(1.2, 3))
}
# Middle and bottom
if(show.header == FALSE & show.tree & (show.confusion | show.levels | show.roc)) {
show.top <- FALSE
show.middle <- TRUE
show.bottom <- TRUE
layout(matrix(1:2, nrow = 2, ncol = 1),
widths = c(6),
heights = c(3, 1.8))
}
# Middle
if(show.header == FALSE & show.tree & (show.confusion == FALSE & show.levels == FALSE & show.roc == FALSE)) {
show.top <- FALSE
show.middle <- TRUE
show.bottom <- FALSE
layout(matrix(1:1, nrow = 1, ncol = 1),
widths = c(6),
heights = c(3))
}
# Bottom
if(show.header == FALSE & show.tree == FALSE) {
show.top <- FALSE
show.middle <- FALSE
show.bottom <- TRUE
nplots <- show.confusion + show.levels + show.roc
layout(matrix(1:nplots, nrow = 1, ncol = nplots),
widths = c(3 * nplots),
heights = c(3))
}
}
# -------------------------
# Setup data
# --------------------------
{
# Extract important parameters from x
goal <- x$params$goal
if(is.null(decision.labels)) {
if(("decision.labels" %in% names(x$params))) {
decision.labels <- x$params$decision.labels
} else {decision.labels <- c(0, 1)}
}
if(is.null(main)) {
if(("main" %in% names(x$params))) {
if(is.null(x$params$main)) {
if(show.header) {
main <- "Data"
} else {main <- ""}
} else {
main <- x$params$main
}
} else {
if(class(data) == "character") {
if(data == "train") {main <- "Data (Training)"}
if(data == "test") {main <- "Data (Testing)"}
}
if(class(data) == "data.frame") {main <- "Test Data"}
}
}
# Check for problems and depreciated arguments
{
if(is.null(which.tree) == FALSE) {
message("The which.tree argument is depreciated and is now just called tree. Please use tree from now on to avoid this message.")
tree <- which.tree
}
if(class(x) != "FFTrees") {
stop("You did not include a valid FFTrees class object or specify the tree directly with level.names, level.classes (etc.). Either create a valid FFTrees object with FFTrees() or specify the tree directly.")
}
if(tree == "best.test" & is.null(x$tree.stats$test)) {
print("You wanted to plot the best test tree (tree = 'best.test') but there were no test data, I'll plot the best training tree instead")
tree <- "best.train"
}
if(is.numeric(tree) & (tree %in% 1:x$trees$n) == F) {
stop(paste("You asked for a tree that does not exist. This object has", x$trees$n, "trees"))
}
if(class(data) == "character") {
if(data == "test" & is.null(x$trees$results$test$stats)) {stop("You asked to plot the test data but there are no test data in the FFTrees object")}
}
}
# DEFINE PLOTTING TREE
if(tree == "best.train") {
tree <- x$trees$best$train
}
if(tree == "best.test") {
tree <- x$trees$best$test
}
# DEFINE CRITICAL OBJECTS
lr.stats <- NULL
cart.stats <- NULL
rf.stats <- NULL
svm.stats <- NULL
decision.v <- x$trees$results[[data]]$decisions[,tree]
tree.stats <- x$trees$results[[data]]$stats
level.stats <- x$trees$results[[data]]$level_stats[x$trees$results[[data]]$level_stats$tree == tree,]
if(comp == TRUE) {
if(is.null(x$comp$lr$results) == FALSE) {
lr.stats <- data.frame("sens" = x$comp$lr$results[[data]]$sens,
"spec" = x$comp$lr$results[[data]]$spec)
}
if(is.null(x$comp$cart$results) == FALSE) {
lcart.stats <- data.frame("sens" = x$comp$cart$results[[data]]$sens,
"spec" = x$comp$cart$results[[data]]$spec)
}
if(is.null(x$comp$rf$results) == FALSE) {
rf.stats <- data.frame("sens" = x$comp$rf$results[[data]]$sens,
"spec" = x$comp$rf$results[[data]]$spec)
}
if(is.null(x$comp$svm$results) == FALSE) {
svm.stats <- data.frame("sens" = x$comp$svm$results[[data]]$sens,
"spec" = x$comp$svm$results[[data]]$spec)
}
}
n.exemplars <- nrow(x$data[[data]])
n.pos <- sum(x$data[[data]][[x$metadata$criterion_name]])
n.neg <- sum(x$data[[data]][[x$metadata$criterion_name]] == FALSE)
mcu <- x$trees$results[[data]]$stats$mcu[tree]
crit.br <- mean(x$data[[data]][[x$metadata$criterion_name]])
final.stats <- tree.stats[tree,]
# ADD LEVEL STATISTICS
n.levels <- nrow(level.stats)
# Add marginal classification statistics to level.stats
level.stats[c("hi.m", "mi.m", "fa.m", "cr.m")] <- NA
for(i in 1:n.levels) {
if(i == 1) {
level.stats[1, c("hi.m", "mi.m", "fa.m", "cr.m")] <- level.stats[1, c("hi", "mi", "fa", "cr")]
}
if(i > 1) {
level.stats[i, c("hi.m", "mi.m", "fa.m", "cr.m")] <- level.stats[i, c("hi", "mi", "fa", "cr")] - level.stats[i - 1, c("hi", "mi", "fa", "cr")]
}
}
}
# -------------------------
# Define plotting parameters
# --------------------------
{
{
# Panels
panel.line.lwd <- 1
panel.line.col <- gray(0)
panel.line.lty <- 1
# Some general parameters
ball.col = c(gray(0), gray(0))
ball.bg = c(gray(1), gray(1))
ball.pch = c(21, 24)
ball.cex = c(1, 1)
error.col <- "red"
correct.col <- "green"
max.label.length <- 100
def.par <- par(no.readonly = TRUE)
ball.box.width <- 10
label.box.height <- 2
label.box.width <- 5
# Define cue labels
{
if(is.null(cue.labels)) {
cue.labels <- level.stats$cue
}
# Trim labels
cue.labels <- strtrim(cue.labels, max.label.length)
}
# Node Segments
segment.lty <- 1
segment.lwd <- 1
continue.segment.lwd <- 1
continue.segment.lty <- 1
exit.segment.lwd <- 1
exit.segment.lty <- 1
decision.node.cex <- 4
exit.node.cex <- 4
panel.title.cex <- 2
# Classification table
classtable.lwd <- 1
# ROC
roc.lwd <- 1
roc.border.col <- gray(0)
# Label sizes
{
# Set cue label size
if(is.null(cue.cex)) {
cue.cex <- c(1.5, 1.5, 1.25, 1, 1, 1)
} else {
if(length(cue.cex) < 6) {cue.cex <- rep(cue.cex, length.out = 6)}
}
# Set break label size
if(is.null(threshold.cex)) {
threshold.cex <- c(1.5, 1.5, 1.25, 1, 1, 1)
} else {
if(length(threshold.cex) < 6) {threshold.cex <- rep(threshold.cex, length.out = 6)}
}
}
if(show.top & show.middle & show.bottom) {
plotting.parameters.df <- data.frame(
n.levels = 1:6,
plot.height = c(10, 12, 15, 19, 23, 27),
plot.width = c(14, 16, 20, 24, 28, 34),
label.box.text.cex = cue.cex,
break.label.cex = threshold.cex
)
} else if (show.top == FALSE & show.middle & show.bottom == FALSE) {
plotting.parameters.df <- data.frame(
n.levels = 1:6,
plot.height = c(10, 12, 15, 19, 23, 25),
plot.width = c(14, 16, 20, 24, 28, 32) * 1,
label.box.text.cex = cue.cex,
break.label.cex = threshold.cex
)
} else {
plotting.parameters.df <- data.frame(
n.levels = 1:6,
plot.height = c(10, 12, 15, 19, 23, 25),
plot.width = c(14, 16, 20, 24, 28, 32) * 1,
label.box.text.cex = cue.cex,
break.label.cex = threshold.cex
)
}
if(n.levels < 6) {
label.box.text.cex <- plotting.parameters.df$label.box.text.cex[n.levels]
break.label.cex <- plotting.parameters.df$break.label.cex[n.levels]
plot.height <- plotting.parameters.df$plot.height[n.levels]
plot.width <- plotting.parameters.df$plot.width[n.levels]
}
if(n.levels >= 6) {
label.box.text.cex <- plotting.parameters.df$label.box.text.cex[6]
break.label.cex <- plotting.parameters.df$break.label.cex[6]
plot.height <- plotting.parameters.df$plot.height[6]
plot.width <- plotting.parameters.df$plot.width[6]
}
# Colors
exit.node.bg <- "white"
error.colfun <- circlize::colorRamp2(c(0, 50, 100),
colors = c("white", "red", "black"))
correct.colfun <- circlize::colorRamp2(c(0, 50, 100),
colors = c("white", "green", "black"))
error.bg <- yarrr::transparent(error.colfun(35), .2)
error.border <- yarrr::transparent(error.colfun(65), .1)
correct.bg <- yarrr::transparent(correct.colfun(35), .2)
correct.border <- yarrr::transparent(correct.colfun(65), .1)
max.cex <- 6
min.cex <- 1
exit.node.pch <- 21
decision.node.pch <- NA_integer_
# balls
ball.loc <- "variable"
if(n.levels == 3) {ball.box.width <- 14}
if(n.levels == 4) {ball.box.width <- 18}
ball.box.height <- 2.5
ball.box.horiz.shift <- 10
ball.box.vert.shift <- -1
ball.box.max.shift.p <- .9
ball.box.min.shift.p <- .4
ball.box.fixed.x.shift <- c(ball.box.min.shift.p * plot.width, ball.box.max.shift.p * plot.width)
# Determine N per ball
if(is.null(n.per.icon)) {
max.n.side <- max(c(n.pos, n.neg))
i <- max.n.side / c(1, 5, 10^(1:10))
i[i > 50] <- 0
n.per.icon <- c(1, 5, 10^(1:10))[which(i == max(i))]
}
noise.ball.pch <- ball.pch[1]
signal.ball.pch <- ball.pch[2]
noise.ball.col <- ball.col[1]
signal.ball.col <- ball.col[2]
noise.ball.bg <- ball.bg[1]
signal.ball.bg <- ball.bg[2]
# arrows
arrow.lty <- 1
arrow.lwd <- 1
arrow.length <- 2.5
arrow.head.length <- .08
arrow.col <- gray(0)
# Final stats
spec.circle.x <- .4
dprime.circle.x <- .5
sens.circle.x <- .6
stat.circle.y <- .3
sens.circle.col <- "green"
spec.circle.col <- "red"
dprime.circle.col <- "blue"
stat.outer.circle.col <- gray(.5)
}
# add.balls.fun() adds balls to the plot
add.balls.fun <- function(x.lim = c(-10, 0),
y.lim = c(-2, 0),
n.vec = c(20, 10),
pch.vec = c(21, 21),
ball.cex = 1,
bg.vec = ball.bg,
col.vec = ball.col,
ball.lwd = .7,
freq.text = TRUE,
freq.text.cex = 1.2,
upper.text = "",
upper.text.cex = 1,
upper.text.adj = 0,
rev.order = F,
box.col = NULL,
box.bg = NULL,
n.per.icon = NULL
) {
# Add box
if(is.null(box.col) == FALSE | is.null(box.bg) == FALSE) {
rect(x.lim[1],
y.lim[1],
x.lim[2],
y.lim[2],
col = box.bg,
border = box.col)
}
# add upper text
text(mean(x.lim), y.lim[2] + upper.text.adj,
label = upper.text, cex = upper.text.cex
)
a.n <- n.vec[1]
b.n <- n.vec[2]
a.p <- n.vec[1] / sum(n.vec)
box.x.center <- sum(x.lim) / 2
box.y.center <- sum(y.lim) / 2
box.x.width <- x.lim[2] - x.lim[1]
# Determine cases per ball
if(is.null(n.per.icon)) {
max.n.side <- max(c(a.n, b.n))
i <- max.n.side / c(1, 5, 10, 50, 100, 1000, 10000)
i[i > 50] <- 0
n.per.icon <- c(1, 5, 10, 50, 100, 1000, 10000)[which(i == max(i))]
}
# Determine general ball locations
a.balls <- ceiling(a.n / n.per.icon)
b.balls <- ceiling(b.n / n.per.icon)
n.balls <- a.balls + b.balls
a.ball.x.loc <- 0
a.ball.y.loc <- 0
b.ball.x.loc <- 0
b.ball.y.loc <- 0
if(a.balls > 0) {
a.ball.x.loc <- rep(-1:-10, each = 5, length.out = 50)[1:a.balls]
a.ball.y.loc <- rep(1:5, times = 10, length.out = 50)[1:a.balls]
a.ball.x.loc <- a.ball.x.loc * (x.lim[2] - box.x.center) / 10 + box.x.center
a.ball.y.loc <- a.ball.y.loc * (y.lim[2] - y.lim[1]) / 5 + y.lim[1]
}
if(b.balls > 0) {
b.ball.x.loc <- rep(1:10, each = 5, length.out = 50)[1:b.balls]
b.ball.y.loc <- rep(1:5, times = 10, length.out = 50)[1:b.balls]
b.ball.x.loc <- b.ball.x.loc * (x.lim[2] - box.x.center) / 10 + box.x.center
b.ball.y.loc <- b.ball.y.loc * (y.lim[2] - y.lim[1]) / 5 + y.lim[1]
}
# if(rev.order) {
#
# x <- b.ball.x.loc
# y <- b.ball.y.loc
#
# b.ball.x.loc <- a.x.loc
# b.ball.y.loc <- a.y.loc
#
# a.ball.x.loc <- x
# a.ball.y.loc <- y
#
# }
# Add frequency text
if(freq.text) {
text(box.x.center, y.lim[1] - 1 * (y.lim[2] - y.lim[1]) / 5, prettyNum(b.n, big.mark = ","), pos = 4, cex = freq.text.cex)
text(box.x.center, y.lim[1] - 1 * (y.lim[2] - y.lim[1]) / 5, prettyNum(a.n, big.mark = ","), pos = 2, cex = freq.text.cex)
}
# Draw balls
# Noise
suppressWarnings(if(a.balls > 0) {
points(x = a.ball.x.loc,
y = a.ball.y.loc,
pch = pch.vec[1],
bg = bg.vec[1],
col = col.vec[1],
cex = ball.cex,
lwd = ball.lwd)
}
)
# Signal
suppressWarnings(if(b.balls > 0) {
points(x = b.ball.x.loc,
y = b.ball.y.loc,
pch = pch.vec[2],
bg = bg.vec[2],
col = col.vec[2],
cex = ball.cex,
lwd = ball.lwd
)
}
)
}
label.cex.fun <- function(i, label.box.text.cex = 2) {
i <- nchar(i)
label.box.text.cex * i ^ -.25
}
}
# -------------------------
# 1: Initial Frequencies
# --------------------------
if(show.top) {
par(mar = c(0, 0, 1, 0))
plot(1, xlim = c(0, 1), ylim = c(0, 1), bty = "n", type = "n",
xlab = "", ylab = "", yaxt = "n", xaxt = "n")
if(hlines) {
segments(0, .95, 1, .95, col = panel.line.col, lwd = panel.line.lwd, lty = panel.line.lty)
}
rect(.33, .8, .67, 1.2, col = "white", border = NA)
text(x = .5, y = .95, main, cex = panel.title.cex)
text(x = .5, y = .80, paste("N = ", prettyNum(n.exemplars, big.mark = ","), "", sep = ""), cex = 1.25)
n.trueneg <- with(final.stats, cr + fa)
n.truepos <- with(final.stats, hi + mi)
text(.5, .65, paste(decision.labels[1], sep = ""), pos = 2, cex = 1.2, adj = 1)
text(.5, .65, paste(decision.labels[2], sep = ""), pos = 4, cex = 1.2, adj = 0)
#points(.9, .8, pch = 1, cex = 1.2)
#text(.9, .8, labels = paste(" = ", n.per.icon, " cases", sep = ""), pos = 4)
# Show ball examples
par(xpd = T)
add.balls.fun(x.lim = c(.35, .65),
y.lim = c(.1, .5),
n.vec = c(final.stats$fa + final.stats$cr, final.stats$hi + final.stats$mi),
pch.vec = c(noise.ball.pch, signal.ball.pch),
bg.vec = c(noise.ball.bg, signal.ball.bg),
col.vec = c(noise.ball.col, signal.ball.col),
ball.cex = ball.cex,
upper.text.adj = 2,
n.per.icon = n.per.icon
)
par(xpd = F)
# Add p.signal and p.noise levels
signal.p <- mean(x$data[[data]][[x$metadata$criterion_name]])
noise.p <- 1 - signal.p
p.rect.ylim <- c(.1, .6)
# p.signal level
text(x = .8, y = p.rect.ylim[2],
labels = paste("p(", decision.labels[2], ")", sep = ""),
pos = 3, cex = 1.2)
#Filling
rect(.775, p.rect.ylim[1],
.825, p.rect.ylim[1] + signal.p * diff(p.rect.ylim),
col = gray(.5, .25), border = NA)
# Filltop
segments(.775, p.rect.ylim[1] + signal.p * diff(p.rect.ylim),
.825, p.rect.ylim[1] + signal.p * diff(p.rect.ylim),
lwd = 1)
#Outline
rect(.775, p.rect.ylim[1],
.825, p.rect.ylim[2], lwd = 1)
if(signal.p < .0001) {signal.p.text <- "<1%"} else {
signal.p.text <- paste(round(signal.p * 100, 0), "%", sep = "")
}
text(.825, p.rect.ylim[1] + signal.p * diff(p.rect.ylim),
labels = signal.p.text,
pos = 4, cex = 1.2)
#p.noise level
text(x = .2, y = p.rect.ylim[2],
labels = paste("p(", decision.labels[1], ")", sep = ""),
pos = 3, cex = 1.2)
rect(.175, p.rect.ylim[1], .225, p.rect.ylim[1] + noise.p * diff(p.rect.ylim),
col = gray(.5, .25), border = NA)
# Filltop
segments(.175, p.rect.ylim[1] + noise.p * diff(p.rect.ylim),
.225, p.rect.ylim[1] + noise.p * diff(p.rect.ylim),
lwd = 1)
# outline
rect(.175, p.rect.ylim[1], .225, p.rect.ylim[2],
lwd = 1)
if(noise.p < .0001) {noise.p.text <- "<0.01%"} else {
noise.p.text <- paste(round(noise.p * 100, 0), "%", sep = "")
}
text(.175, p.rect.ylim[1] + noise.p * diff(p.rect.ylim),
labels = noise.p.text,
pos = 2, cex = 1.2)
}
# -------------------------
# 2. TREE
# --------------------------
if(show.middle) {
if(show.top == FALSE & show.bottom == FALSE) {
par(mar = c(3, 3, 3, 3) + .1)
} else {par(mar = c(0, 0, 0, 0))
}
# Setup plotting space
plot(1,
xlim = c(-plot.width, plot.width),
ylim = c(-plot.height, 0),
type = "n", bty = "n",
xaxt = "n", yaxt = "n",
ylab = "", xlab = ""
)
# Add frame
par(xpd = TRUE)
if(show.top | show.bottom) {
if(hlines) {
segments(-plot.width, 0, - plot.width * .3, 0, col = panel.line.col, lwd = panel.line.lwd, lty = panel.line.lty)
segments(plot.width, 0, plot.width * .3, 0, col = panel.line.col, lwd = panel.line.lwd, lty = panel.line.lty)
}
if(is.null(label.tree)) {label.tree <- paste("FFT #", tree, " (of ", x$trees$n, ")", sep = "")}
text(x = 0, y = 0,
label.tree,
cex = panel.title.cex)
}
if(show.top == FALSE & show.bottom == FALSE) {
if(is.null(main) & is.null(x$params$main)) {main <- ""}
mtext(text = main, side = 3, cex = 2)
}
par(xpd = FALSE)
# Create Noise and Signal panels
if(show.iconguide) {
# Noise Balls
points(c(- plot.width * .7, - plot.width * .5),
c(-plot.height * .125, -plot.height * .125),
pch = c(noise.ball.pch, signal.ball.pch),
bg = c(correct.bg, error.bg),
col = c(correct.border, error.border),
cex = ball.cex * 1.5
)
text(c(- plot.width * .7, - plot.width * .5),
c(-plot.height * .125, -plot.height * .125),
labels = c("Correct\nRejection", "Miss"),
pos = c(2, 4), offset = 1)
# Noise Panel
text(- plot.width * .6, -plot.height * .05,
paste("Decide ", decision.labels[1], sep = ""),
cex = 1.2, font = 3
)
# Signal panel
text(plot.width * .6, -plot.height * .05,
paste("Decide ", decision.labels[2], sep = ""),
cex = 1.2, font = 3
)
points(c(plot.width * .5, plot.width * .7),
c(-plot.height * .125, -plot.height * .125),
pch = c(noise.ball.pch, signal.ball.pch),
bg = c(error.bg, correct.bg),
col = c(error.border, correct.border),
cex = ball.cex * 1.5
)
text(c(plot.width * .5, plot.width * .7),
c(-plot.height * .125, -plot.height * .125),
labels = c("False\nAlarm", "Hit"),
pos = c(2, 4), offset = 1)
}
# Set initial subplot center
subplot.center <- c(0, -4)
# Loop over levels
for(level.i in 1:min(c(n.levels, 6))) {
current.cue <- cue.labels[level.i]
# Get stats for current level
{
hi.i <- level.stats$hi.m[level.i]
mi.i <- level.stats$mi.m[level.i]
fa.i <- level.stats$fa.m[level.i]
cr.i <- level.stats$cr.m[level.i]
}
# If level.i == 1, draw top textbox
{
if(level.i == 1) {
rect(subplot.center[1] - label.box.width / 2,
subplot.center[2] + 2 - label.box.height / 2,
subplot.center[1] + label.box.width / 2,
subplot.center[2] + 2 + label.box.height / 2,
col = "white",
border = "black"
)
points(x = subplot.center[1],
y = subplot.center[2] + 2,
cex = decision.node.cex,
pch = decision.node.pch
)
text(x = subplot.center[1],
y = subplot.center[2] + 2,
labels = current.cue,
cex = label.box.text.cex#label.cex.fun(current.cue, label.box.text.cex = label.box.text.cex)
)
}
}
# -----------------------
# Left (Noise) Classification / New Level
# -----------------------
{
# Exit node on left
if(level.stats$exit[level.i] %in% c(0, .5) | paste(level.stats$exit[level.i]) %in% c("0", ".5")) {
segments(subplot.center[1],
subplot.center[2] + 1,
subplot.center[1] - 2,
subplot.center[2] - 2,
lty = segment.lty,
lwd = segment.lwd
)
arrows(x0 = subplot.center[1] - 2,
y0 = subplot.center[2] - 2,
x1 = subplot.center[1] - 2 - arrow.length,
y1 = subplot.center[2] - 2,
lty = arrow.lty,
lwd = arrow.lwd,
col = arrow.col,
length = arrow.head.length
)
# Decision text
if(decision.cex > 0) {
text(x = subplot.center[1] - 2 - arrow.length * .7,
y = subplot.center[2] - 2.2,
labels = decision.labels[1],
pos = 1, font = 3, cex = decision.cex)
}
if(ball.loc == "fixed") {
ball.x.lim <- c(-max(ball.box.fixed.x.shift), -min(ball.box.fixed.x.shift))
ball.y.lim <- c(subplot.center[2] + ball.box.vert.shift - ball.box.height / 2,
subplot.center[2] + ball.box.vert.shift + ball.box.height / 2)
}
if(ball.loc == "variable") {
ball.x.lim <- c(subplot.center[1] - ball.box.horiz.shift - ball.box.width / 2,
subplot.center[1] - ball.box.horiz.shift + ball.box.width / 2)
ball.y.lim <- c(subplot.center[2] + ball.box.vert.shift - ball.box.height / 2,
subplot.center[2] + ball.box.vert.shift + ball.box.height / 2)
}
if(max(c(cr.i, mi.i), na.rm = TRUE) > 0 & show.icons == TRUE) {
add.balls.fun(x.lim = ball.x.lim,
y.lim = ball.y.lim,
n.vec = c(cr.i, mi.i),
pch.vec = c(noise.ball.pch, signal.ball.pch),
# bg.vec = c(noise.ball.bg, signal.ball.bg),
bg.vec = c(correct.bg, error.bg),
col.vec = c(correct.border, error.border),
freq.text = TRUE,
n.per.icon = n.per.icon,
ball.cex = ball.cex
)
}
# level break label
pos.direction.symbol <- c("<=", "<", "=", "!=", ">", ">=")[which(level.stats$direction[level.i] == c(">", ">=", "!=", "=", "<=", "<"))]
neg.direction.symbol <- c("<=", "<", "=", "!=", ">", ">=")[which(level.stats$direction[level.i] == c("<=", "<", "=", "!=", ">", ">="))]
text.outline(x = subplot.center[1] - 1,
y = subplot.center[2],
labels = paste(pos.direction.symbol, " ", level.stats$threshold[level.i], sep = ""),
pos = 2, cex = break.label.cex, r = .1
)
points(x = subplot.center[1] - 2,
y = subplot.center[2] - 2,
pch = exit.node.pch,
cex = exit.node.cex,
bg = exit.node.bg
)
text(x = subplot.center[1] - 2,
y = subplot.center[2] - 2,
labels = substr(decision.labels[1], 1, 1)
)
}
# New level on left
if(level.stats$exit[level.i] %in% c(1) | paste(level.stats$exit[level.i]) %in% c("1")) {
segments(subplot.center[1],
subplot.center[2] + 1,
subplot.center[1] - 2,
subplot.center[2] - 2,
lty = segment.lty,
lwd = segment.lwd
)
rect(subplot.center[1] - 2 - label.box.width / 2,
subplot.center[2] - 2 - label.box.height / 2,
subplot.center[1] - 2 + label.box.width / 2,
subplot.center[2] - 2 + label.box.height / 2,
col = "white",
border = "black"
)
if(level.i < 6) {text(x = subplot.center[1] - 2,
y = subplot.center[2] - 2,
labels = cue.labels[level.i + 1],
cex = label.box.text.cex
)} else {
text(x = subplot.center[1] - 2,
y = subplot.center[2] - 2,
labels = paste0("+ ", n.levels - 6, " More"),
cex = label.box.text.cex,
font = 3
)
}
}
}
# -----------------------
# Right (Signal) Classification / New Level
# -----------------------
{
# Exit node on right
if(level.stats$exit[level.i] %in% c(1, .5) | paste(level.stats$exit[level.i]) %in% c("1", ".5")) {
segments(subplot.center[1],
subplot.center[2] + 1,
subplot.center[1] + 2,
subplot.center[2] - 2,
lty = segment.lty,
lwd = segment.lwd
)
arrows(x0 = subplot.center[1] + 2,
y0 = subplot.center[2] - 2,
x1 = subplot.center[1] + 2 + arrow.length,
y1 = subplot.center[2] - 2,
lty = arrow.lty,
lwd = arrow.lwd,
col = arrow.col,
length = arrow.head.length
)
# Decision text
if(decision.cex > 0) {
text(x = subplot.center[1] + 2 + arrow.length * .7,
y = subplot.center[2] - 2.2,
labels = decision.labels[2],
pos = 1, font = 3, cex = decision.cex)
}
if(ball.loc == "fixed") {
ball.x.lim <- c(min(ball.box.fixed.x.shift), max(ball.box.fixed.x.shift))
ball.y.lim <- c(subplot.center[2] + ball.box.vert.shift - ball.box.height / 2,
subplot.center[2] + ball.box.vert.shift + ball.box.height / 2)
}
if(ball.loc == "variable") {
ball.x.lim <- c(subplot.center[1] + ball.box.horiz.shift - ball.box.width / 2,
subplot.center[1] + ball.box.horiz.shift + ball.box.width / 2)
ball.y.lim <- c(subplot.center[2] + ball.box.vert.shift - ball.box.height / 2,
subplot.center[2] + ball.box.vert.shift + ball.box.height / 2)
}
if(max(c(fa.i, hi.i), na.rm = TRUE) > 0 & show.icons == TRUE) {
add.balls.fun(x.lim = ball.x.lim,
y.lim = ball.y.lim,
n.vec = c(fa.i, hi.i),
pch.vec = c(noise.ball.pch, signal.ball.pch),
# bg.vec = c(noise.ball.bg, signal.ball.bg),
bg.vec = c(error.bg, correct.bg),
col.vec = c(error.border, correct.border),
freq.text = TRUE,
n.per.icon = n.per.icon,
ball.cex = ball.cex
)
}
# level break label
dir.symbols <- c("<=", "<", "=", "!=", ">", ">=")
pos.direction.symbol <- dir.symbols[which(level.stats$direction[level.i] == c("<=", "<", "=", "!=", ">", ">="))]
neg.direction.symbol <- dir.symbols[which(level.stats$direction[level.i] == rev(c("<=", "<", "=", "!=", ">", ">=")))]
text.outline(subplot.center[1] + 1,
subplot.center[2],
labels = paste(pos.direction.symbol, " ", level.stats$threshold[level.i], sep = ""),
pos = 4, cex = break.label.cex, r = .1
)
points(x = subplot.center[1] + 2,
y = subplot.center[2] - 2,
pch = exit.node.pch,
cex = exit.node.cex,
bg = exit.node.bg
)
text(x = subplot.center[1] + 2,
y = subplot.center[2] - 2,
labels = substr(decision.labels[2], 1, 1)
)
}
# New level on right
if(level.stats$exit[level.i] %in% 0 | paste(level.stats$exit[level.i]) %in% c("0")) {
segments(subplot.center[1],
subplot.center[2] + 1,
subplot.center[1] + 2,
subplot.center[2] - 2,
lty = segment.lty,
lwd = segment.lwd
)
if(level.i < 6) {
rect(subplot.center[1] + 2 - label.box.width / 2,
subplot.center[2] - 2 - label.box.height / 2,
subplot.center[1] + 2 + label.box.width / 2,
subplot.center[2] - 2 + label.box.height / 2,
col = "white",
border = "black")
text(x = subplot.center[1] + 2,
y = subplot.center[2] - 2,
labels = cue.labels[level.i + 1],
cex = label.box.text.cex)
} else {
rect(subplot.center[1] + 2 - label.box.width / 2,
subplot.center[2] - 2 - label.box.height / 2,
subplot.center[1] + 2 + label.box.width / 2,
subplot.center[2] - 2 + label.box.height / 2,
col = "white",
border = "black", lty = 2)
text(x = subplot.center[1] + 2,
y = subplot.center[2] - 2,
labels = paste0("+ ", n.levels - 6, " More"),
cex = label.box.text.cex,
font = 3
)
}
}
}
# -----------------------
# Update plot center
# -----------------------
{
if(level.stats$exit[level.i] == 0) {
subplot.center <- c(subplot.center[1] + 2,
subplot.center[2] - 4)
}
if(level.stats$exit[level.i] == 1) {
subplot.center <- c(subplot.center[1] - 2,
subplot.center[2] - 4)
}
}
}
}
# -----------------------
# 3. CUMULATIVE PERFORMANCE
# -----------------------
if(show.bottom == TRUE) {
{
# OBTAIN FINAL STATISTICS
fft.sens.vec <- tree.stats$sens
fft.spec.vec <- tree.stats$spec
# General plotting space
{
# PLOTTING PARAMETERS
header.y.loc <- 1.0
subheader.y.loc <- .925
header.cex <- 1.1
subheader.cex <- .9
par(mar = c(0, 0, 2, 0))
plot(1, xlim = c(0, 1), ylim = c(0, 1),
bty = "n", type = "n",
xlab = "", ylab = "",
yaxt = "n", xaxt = "n")
par(xpd = T)
if(hlines) {
segments(0, 1.1, 1, 1.1, col = panel.line.col, lwd = panel.line.lwd, lty = panel.line.lty)
rect(.25, 1, .75, 1.2, col = "white", border = NA)
}
if(is.null(label.performance)) {
if(data == "train") {label.performance <- "Accuracy (Training)"}
if(data == "test") {label.performance <- "Accuracy (Testing)"}
}
text(.5, 1.1, label.performance, cex = panel.title.cex)
par(xpd = FALSE)
pretty.dec <- function(x) {return(paste(round(x, 2) * 100, sep = ""))}
level.max.height <- .65
level.width <- .05
level.center.y <- .45
#level.bottom <- .1
level.bottom <- level.center.y - level.max.height / 2
level.top <- level.center.y + level.max.height / 2
lloc <- data.frame(
element = c("classtable", "mcu", "pci", "sens", "spec", "acc", "bacc", "roc"),
long.name = c("Classification Table", "mcu", "pci", "sens", "spec", "acc", "bacc", "ROC"),
center.x = c(.18, seq(.35, .65, length.out = 6), .85),
center.y = rep(level.center.y, 8),
width = c(.2, rep(level.width, 6), .2),
height = c(.65, rep(level.max.height, 6), .65),
value = c(NA,
abs(final.stats$mcu - 5) / (abs(1 - 5)),
final.stats$pci, final.stats$sens, final.stats$spec, with(final.stats, (cr + hi) / n), final.stats$bacc, NA),
value.name = c(NA, round(final.stats$mcu, 1), pretty.dec(final.stats$pci), pretty.dec(final.stats$sens), pretty.dec(final.stats$spec), pretty.dec(final.stats$acc),
pretty.dec(final.stats$bacc), NA
)
)
}
# Classification table
if(show.confusion) {
final.classtable.x.loc <- c(lloc$center.x[lloc$element == "classtable"] - lloc$width[lloc$element == "classtable"] / 2, lloc$center.x[lloc$element == "classtable"] + lloc$width[lloc$element == "classtable"] / 2)
final.classtable.y.loc <- c(lloc$center.y[lloc$element == "classtable"] - lloc$height[lloc$element == "classtable"] / 2, lloc$center.y[lloc$element == "classtable"] + lloc$height[lloc$element == "classtable"] / 2)
rect(final.classtable.x.loc[1], final.classtable.y.loc[1],
final.classtable.x.loc[2], final.classtable.y.loc[2],
lwd = classtable.lwd
)
segments(mean(final.classtable.x.loc), final.classtable.y.loc[1], mean(final.classtable.x.loc), final.classtable.y.loc[2], col = gray(0), lwd = classtable.lwd)
segments(final.classtable.x.loc[1], mean(final.classtable.y.loc), final.classtable.x.loc[2], mean(final.classtable.y.loc), col = gray(0), lwd = classtable.lwd)
# Column titles
text(x = mean(mean(final.classtable.x.loc)),
y = header.y.loc,
"Truth", pos = 1, cex = header.cex)
text(x = final.classtable.x.loc[1] + .25 * diff(final.classtable.x.loc),
y = subheader.y.loc, pos = 1, cex = subheader.cex,
decision.labels[2])
text(x = final.classtable.x.loc[1] + .75 * diff(final.classtable.x.loc),
y = subheader.y.loc, pos = 1, cex = subheader.cex,
decision.labels[1])
# Row titles
text(x = final.classtable.x.loc[1] - .01,
y = final.classtable.y.loc[1] + .75 * diff(final.classtable.y.loc), cex = subheader.cex,
decision.labels[2], adj = 1)
text(x = final.classtable.x.loc[1] - .01,
y = final.classtable.y.loc[1] + .25 * diff(final.classtable.y.loc), cex = subheader.cex,
decision.labels[1], adj = 1)
text(x = final.classtable.x.loc[1] - .065,
y = mean(final.classtable.y.loc), cex = header.cex,
"Decision")
# text(x = final.classtable.x.loc[1] - .05,
# y = mean(final.classtable.y.loc), cex = header.cex,
# "Decision", srt = 90, pos = 3)
# Add final frequencies
text(final.classtable.x.loc[1] + .75 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .25 * diff(final.classtable.y.loc),
prettyNum(final.stats$cr, big.mark = ","), cex = 1.5)
text(final.classtable.x.loc[1] + .25 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .25 * diff(final.classtable.y.loc),
prettyNum(final.stats$mi, big.mark = ","), cex = 1.5)
text(final.classtable.x.loc[1] + .75 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .75 * diff(final.classtable.y.loc),
prettyNum(final.stats$fa, big.mark = ","), cex = 1.5)
text(final.classtable.x.loc[1] + .25 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .75 * diff(final.classtable.y.loc),
prettyNum(final.stats$hi, big.mark = ","), cex = 1.5)
# Add symbols
points(final.classtable.x.loc[1] + .55 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .05 * diff(final.classtable.y.loc),
pch = noise.ball.pch, bg = correct.bg, col = correct.border, cex = ball.cex)
points(final.classtable.x.loc[1] + .05 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .55 * diff(final.classtable.y.loc),
pch = signal.ball.pch, bg = correct.bg, cex = ball.cex, col = correct.border)
points(final.classtable.x.loc[1] + .55 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .55 * diff(final.classtable.y.loc),
pch = noise.ball.pch, bg = error.bg, col = error.border, cex = ball.cex)
points(final.classtable.x.loc[1] + .05 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .05 * diff(final.classtable.y.loc),
pch = signal.ball.pch, bg = error.bg, col = error.border, cex = ball.cex)
# Labels
text(final.classtable.x.loc[1] + .62 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .07 * diff(final.classtable.y.loc),
"cr", cex = 1, font = 3, adj = 0)
text(final.classtable.x.loc[1] + .12 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .07 * diff(final.classtable.y.loc),
"mi", cex = 1, font = 3, adj = 0)
text(final.classtable.x.loc[1] + .62 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .57 * diff(final.classtable.y.loc),
"fa", cex = 1, font = 3, adj = 0)
text(final.classtable.x.loc[1] + .12 * diff(final.classtable.x.loc),
final.classtable.y.loc[1] + .57 * diff(final.classtable.y.loc),
"hi", cex = 1, font = 3, adj = 0)
}
# Levels
if(show.levels) {
if(level.type %in% c("line", "bar")) {
# Color function (taken from colorRamp2 function in circlize package)
col.fun <- circlize::colorRamp2(c(0, .75, 1), c("red", "yellow", "green"), transparency = .1)
add.level.fun <- function(name,
sub = "",
max.val = 1,
min.val = 0,
ok.val = .5,
bottom.text = "",
level.type = "line") {
rect.center.x <- lloc$center.x[lloc$element == name]
rect.center.y <- lloc$center.y[lloc$element == name]
rect.height <- lloc$height[lloc$element == name]
rect.width <- lloc$width[lloc$element == name]
rect.bottom.y <- rect.center.y - rect.height / 2
rect.top.y <- rect.center.y + rect.height / 2
rect.left.x <- rect.center.x - rect.width / 2
rect.right.x <- rect.center.x + rect.width / 2
long.name <- lloc$long.name[lloc$element == name]
value <- lloc$value[lloc$element == name]
value.name <- lloc$value.name[lloc$element == name]
#
# level.col.fun <- circlize::colorRamp2(c(min.val, ok.val, max.val),
# colors = c("firebrick2", "yellow", "green4"),
# transparency = .1)
text(x = rect.center.x,
y = header.y.loc,
labels = long.name,
pos = 1, cex = header.cex)
#
# text.outline(x = rect.center.x,
# y = header.y.loc,
# labels = long.name,
# pos = 1, cex = header.cex, r = .02
# )
value.height <- rect.bottom.y + min(c(1, ((value - min.val) / (max.val - min.val)))) * rect.height
# Add filling
value.s <- min(value / max.val, 1)
delta <- 1
gamma <- .5
value.col.scale <- delta * value.s ^ gamma / (delta * value.s ^ gamma + (1 - value.s) ^ gamma)
# value.col <- gray(1 - value.col.scale * .5)
value.col <- gray(1, .25)
#plot(seq(0, 1, .01), delta * seq(0, 1, .01) ^ gamma / (delta * seq(0, 1, .01) ^ gamma + (1 - seq(0, 1, .01)) ^ gamma))
if(level.type == "bar") {
rect(rect.left.x,
rect.bottom.y,
rect.right.x,
value.height,
#col = level.col.fun(value.s),
col = value.col,
# col = spec.level.fun(lloc$value[lloc$element == name]),
border = "black"
)
text.outline(x = rect.center.x,
y = value.height,
labels = lloc$value.name[lloc$element == name],
cex = 1.5, r = .008, pos = 3
)
# Add level border
# rect(rect.left.x,
# rect.bottom.y,
# rect.right.x,
# rect.top.y,
# border = gray(.5, .5))
}
if(level.type == "line") {
# Stem
segments(rect.center.x,
rect.bottom.y,
rect.center.x,
value.height,
lty = 3)
# Horizontal platform
platform.width <- .02
segments(rect.center.x - platform.width,
value.height,
rect.center.x + platform.width,
value.height)
# Text label
text.outline(x = rect.center.x,
y = value.height,
labels = lloc$value.name[lloc$element == name],
cex = 1.5, r = 0, pos = 3
)
# points(rect.center.x,
# value.height,
# cex = 5.5,
# pch = 21,
# bg = "white",
# col = "black", lwd = .5)
}
# Add subtext
text(x = rect.center.x,
y = rect.center.y - .05,
labels = sub,
cex = .8,
font = 1,
pos = 1)
# Add bottom text
text(x = rect.center.x,
y = rect.bottom.y,
labels = bottom.text,
pos = 1)
}
paste(final.stats$cr, "/", 1, collapse = "")
#Add 100% reference line
# segments(x0 = lloc$center.x[lloc$element == "mcu"] - lloc$width[lloc$element == "mcu"] * .8,
# y0 = level.top,
# x1 = lloc$center.x[lloc$element == "bacc"] + lloc$width[lloc$element == "bacc"] * .8,
# y1 = level.top,
# lty = 3, lwd = .75)
add.level.fun("mcu",
ok.val = .75,
max.val = 1,
min.val = 0,
level.type = level.type) #, sub = paste(c(final.stats$cr, "/", final.stats$cr + final.stats$fa), collapse = ""))
add.level.fun("pci", ok.val = .75, level.type = level.type) #, sub = paste(c(final.stats$cr, "/", final.stats$cr + final.stats$fa), collapse = ""))
# text(lloc$center.x[lloc$element == "pci"],
# lloc$center.y[lloc$element == "pci"],
# labels = paste0("mcu\n", round(mcu, 2)))
add.level.fun("spec", ok.val = .75, level.type = level.type) #, sub = paste(c(final.stats$cr, "/", final.stats$cr + final.stats$fa), collapse = ""))
add.level.fun("sens", ok.val = .75, level.type = level.type) #, sub = paste(c(final.stats$hi, "/", final.stats$hi + final.stats$mi), collapse = ""))
# Min acc
min.acc <- max(crit.br, 1 - crit.br)
add.level.fun("acc", min.val = 0, ok.val = .5, level.type = level.type) #, sub = paste(c(final.stats$hi + final.stats$cr, "/", final.stats$n), collapse = ""))
# Add baseline to acc level
segments(x0 = lloc$center.x[lloc$element == "acc"] - lloc$width[lloc$element == "acc"] / 2,
y0 = (lloc$center.y[lloc$element == "acc"] - lloc$height[lloc$element == "acc"] / 2) + lloc$height[lloc$element == "acc"] * min.acc,
x1 = lloc$center.x[lloc$element == "acc"] + lloc$width[lloc$element == "acc"] / 2,
y1 = (lloc$center.y[lloc$element == "acc"] - lloc$height[lloc$element == "acc"] / 2) + lloc$height[lloc$element == "acc"] * min.acc,
lty = 3)
text(x = lloc$center.x[lloc$element == "acc"],
y =(lloc$center.y[lloc$element == "acc"] - lloc$height[lloc$element == "acc"] / 2) + lloc$height[lloc$element == "acc"] * min.acc,
labels = "BL", pos = 1)
# paste("BL = ", pretty.dec(min.acc), sep = ""), pos = 1)
add.level.fun("bacc", min.val = 0, max.val = 1, ok.val = .5, level.type = level.type)
# baseline
# segments(x0 = mean(lloc$center.x[2]),
# y0 = lloc$center.y[1] - lloc$height[1] / 2,
# x1 = mean(lloc$center.x[7]),
# y1 = lloc$center.y[1] - lloc$height[1] / 2, lend = 1,
# lwd = .5,
# col = gray(0))
}
}
# MiniROC
if(show.roc) {
text(lloc$center.x[lloc$element == "roc"], header.y.loc, "ROC", pos = 1, cex = header.cex)
# text(final.roc.center[1], subheader.y.loc, paste("AUC =", round(final.auc, 2)), pos = 1)
final.roc.x.loc <- c(lloc$center.x[lloc$element == "roc"] - lloc$width[lloc$element == "roc"] / 2,lloc$center.x[lloc$element == "roc"] + lloc$width[lloc$element == "roc"] / 2)
final.roc.y.loc <- c(lloc$center.y[lloc$element == "roc"] - lloc$height[lloc$element == "roc"] / 2,lloc$center.y[lloc$element == "roc"] + lloc$height[lloc$element == "roc"] / 2)
# Plot bg
#
# rect(final.roc.x.loc[1],
# final.roc.y.loc[1],
# final.roc.x.loc[2],
# final.roc.y.loc[2],
# col = gray(1), lwd = .5)
# Gridlines
# # Horizontal
# segments(x0 = rep(final.roc.x.loc[1], 9),
# y0 = seq(final.roc.y.loc[1], final.roc.y.loc[2], length.out = 5)[2:10],
# x1 = rep(final.roc.x.loc[2], 9),
# y1 = seq(final.roc.y.loc[1], final.roc.y.loc[2], length.out = 5)[2:10],
# lty = 1, col = gray(.8), lwd = c(.5), lend = 3
# )
#
# # Vertical
# segments(y0 = rep(final.roc.y.loc[1], 9),
# x0 = seq(final.roc.x.loc[1], final.roc.x.loc[2], length.out = 5)[2:10],
# y1 = rep(final.roc.y.loc[2], 9),
# x1 = seq(final.roc.x.loc[1], final.roc.x.loc[2], length.out = 5)[2:10],
# lty = 1, col = gray(.8), lwd = c(.5), lend = 3
# )
# Plot border
rect(final.roc.x.loc[1],
final.roc.y.loc[1],
final.roc.x.loc[2],
final.roc.y.loc[2],
border = roc.border.col,
lwd = roc.lwd)
# Axis labels
text(c(final.roc.x.loc[1], final.roc.x.loc[2]),
c(final.roc.y.loc[1], final.roc.y.loc[1]) - .05,
labels = c(0, 1))
text(c(final.roc.x.loc[1], final.roc.x.loc[1], final.roc.x.loc[1]) - .02,
c(final.roc.y.loc[1], mean(final.roc.y.loc[1:2]), final.roc.y.loc[2]),
labels = c(0,.5, 1))
text(mean(final.roc.x.loc), final.roc.y.loc[1] - .08, "1 - Specificity (FAR)")
text(final.roc.x.loc[1] - .04, mean(final.roc.y.loc), "Sensitivity (HR)", srt = 90)
# Diagonal
segments(final.roc.x.loc[1],
final.roc.y.loc[1],
final.roc.x.loc[2],
final.roc.y.loc[2],
lty = 3)
label.loc <- c(.1, .3, .5, .7, .9)
## COMPETITIVE ALGORITHMS
if(comp == TRUE) {
# CART
if(is.null(x$comp$cart$results) == FALSE) {
cart.spec <- x$comp$cart$results[[data]]$spec
cart.sens <- x$comp$cart$results[[data]]$sens
points(final.roc.x.loc[1] + (1 - cart.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + cart.sens * lloc$height[lloc$element == "roc"],
pch = 21, cex = 1.75,
col = yarrr::transparent("red", .1),
bg = yarrr::transparent("red", .9), lwd = 1)
points(final.roc.x.loc[1] + (1 - cart.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + cart.sens * lloc$height[lloc$element == "roc"],
pch = "C", cex = .7, col = gray(.2), lwd = 1)
par("xpd" = FALSE)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[4] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5,
col = yarrr::transparent("red", .1),
bg = yarrr::transparent("red", .9))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[4] * lloc$height[lloc$element == "roc"],
pch = "C", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[4] * lloc$height[lloc$element == "roc"],
labels = " CART", adj = 0, cex = .9)
par("xpd" = TRUE)
}
## LR
if(is.null(x$comp$lr$results) == FALSE) {
lr.spec <- x$comp$lr$results[[data]]$spec
lr.sens <- x$comp$lr$results[[data]]$sens
points(final.roc.x.loc[1] + (1 - lr.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + lr.sens * lloc$height[lloc$element == "roc"],
pch = 21, cex = 1.75,
col = yarrr::transparent("blue", .1),
bg = yarrr::transparent("blue", .9))
points(final.roc.x.loc[1] + (1 - lr.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + lr.sens * lloc$height[lloc$element == "roc"],
pch = "L", cex = .7, col = gray(.2))
par("xpd" = F)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[3] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5,
col = yarrr::transparent("blue", .1),
bg = yarrr::transparent("blue", .9))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[3] * lloc$height[lloc$element == "roc"],
pch = "L", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[3] * lloc$height[lloc$element == "roc"],
labels = " LR", adj = 0, cex = .9)
par("xpd" = T)
}
## rf
if(is.null(x$comp$rf$results) == FALSE) {
rf.spec <- x$comp$rf$results[[data]]$spec
rf.sens <- x$comp$rf$results[[data]]$sens
points(final.roc.x.loc[1] + (1 - rf.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + rf.sens * lloc$height[lloc$element == "roc"],
pch = 21, cex = 1.75,
col = yarrr::transparent("purple", .1),
bg = yarrr::transparent("purple", .9), lwd = 1)
points(final.roc.x.loc[1] + (1 - rf.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + rf.sens * lloc$height[lloc$element == "roc"],
pch = "C", cex = .7, col = gray(.2), lwd = 1)
par("xpd" = FALSE)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[2] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5,
col = yarrr::transparent("purple", .1),
bg = yarrr::transparent("purple", .9))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[2] * lloc$height[lloc$element == "roc"],
pch = "R", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[2] * lloc$height[lloc$element == "roc"],
labels = " RF", adj = 0, cex = .9)
par("xpd" = TRUE)
}
## svm
## svm
if(is.null(x$comp$svm$results) == FALSE) {
svm.spec <- x$comp$svm$results[[data]]$spec
svm.sens <- x$comp$svm$results[[data]]$sens
points(final.roc.x.loc[1] + (1 - svm.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + svm.sens * lloc$height[lloc$element == "roc"],
pch = 21, cex = 1.75,
col = yarrr::transparent("orange", .1),
bg = yarrr::transparent("orange", .9), lwd = 1)
points(final.roc.x.loc[1] + (1 - svm.spec) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + svm.sens * lloc$height[lloc$element == "roc"],
pch = "C", cex = .7, col = gray(.2), lwd = 1)
par("xpd" = FALSE)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[1] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5,
col = yarrr::transparent("orange", .1),
bg = yarrr::transparent("orange", .9))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[1] * lloc$height[lloc$element == "roc"],
pch = "S", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[1] * lloc$height[lloc$element == "roc"],
labels = " SVM", adj = 0, cex = .9)
par("xpd" = TRUE)
}
}
## FFT
{
roc.order <- order(fft.spec.vec, decreasing = TRUE)
# roc.order <- 1:x$trees$n
fft.sens.vec.ord <- fft.sens.vec[roc.order]
fft.spec.vec.ord <- fft.spec.vec[roc.order]
# Add segments and points for all trees but tree
if(length(roc.order) > 1) {
segments(final.roc.x.loc[1] + c(0, 1 - fft.spec.vec.ord) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + c(0, fft.sens.vec.ord) * lloc$height[lloc$element == "roc"],
final.roc.x.loc[1] + c(1 - fft.spec.vec.ord, 1) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + c(fft.sens.vec.ord, 1) * lloc$height[lloc$element == "roc"],
lwd = 1,
col = gray(0))
points(final.roc.x.loc[1] + (1 - fft.spec.vec.ord[-(which(roc.order == tree))]) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + fft.sens.vec.ord[-(which(roc.order == tree))] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5, col = yarrr::transparent("green", .3),
bg = yarrr::transparent("white", .1))
text(final.roc.x.loc[1] + (1 - fft.spec.vec.ord[-(which(roc.order == tree))]) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + fft.sens.vec.ord[-(which(roc.order == tree))] * lloc$height[lloc$element == "roc"],
labels = roc.order[which(roc.order != tree)], cex = 1, col = gray(.2))
}
# Add large point for plotted tree
points(final.roc.x.loc[1] + (1 - fft.spec.vec[tree]) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + fft.sens.vec[tree] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 3, col = gray(1), #col = yarrr::transparent("green", .3),
bg = yarrr::transparent("green", .3), lwd = 1)
text(final.roc.x.loc[1] + (1 - fft.spec.vec[tree]) * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + fft.sens.vec[tree] * lloc$height[lloc$element == "roc"],
labels = tree, cex = 1.25, col = gray(.2), font = 2)
# Labels
if(comp == TRUE & any(is.null(x$comp$lr$results) == FALSE,
is.null(x$comp$cart$results) == FALSE,
is.null(x$comp$svm$results) == FALSE,
is.null(x$comp$rf$results) == FALSE
)) {
par("xpd" = FALSE)
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[5] * lloc$height[lloc$element == "roc"],
pch = 21, cex = 2.5, col = yarrr::transparent("green", .3),
bg = yarrr::transparent("green", .7))
points(final.roc.x.loc[1] + 1.1 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[5] * lloc$height[lloc$element == "roc"],
pch = "#", cex = .9, col = gray(.2))
text(final.roc.x.loc[1] + 1.13 * lloc$width[lloc$element == "roc"],
final.roc.y.loc[1] + label.loc[5] * lloc$height[lloc$element == "roc"],
labels = " FFT", adj = 0, cex = .9)
par("xpd" = TRUE)
}
}
}
}
}
# Reset plotting space
par(mfrow = c(1, 1))
par(mar = c(5, 4, 4, 1) + .1)
}
}
|
if(FALSE) {
getIfValue(findIfInFun(foo)[[1]])
i = findIfInFun(foo)[[1]]
}
foo =
function(x)
{
y = 2
if(length(x))
y = y + 3
else
y = 10
y
}
foo2 =
function(x)
{
y = 2L
names(y) = "a"
y[2] = 11 # changes to numeric.
if(length(x))
y = y + 3
else
y = 10
y
}
bar =
function(x)
{
if(length(x))
x + 3
else
10
}
| /explorations/ifAnalysis/egIf.R | no_license | duncantl/CodeAnalysis | R | false | false | 415 | r | if(FALSE) {
getIfValue(findIfInFun(foo)[[1]])
i = findIfInFun(foo)[[1]]
}
foo =
function(x)
{
y = 2
if(length(x))
y = y + 3
else
y = 10
y
}
foo2 =
function(x)
{
y = 2L
names(y) = "a"
y[2] = 11 # changes to numeric.
if(length(x))
y = y + 3
else
y = 10
y
}
bar =
function(x)
{
if(length(x))
x + 3
else
10
}
|
# ---
# title: "1"
# objective: "Function as tangents"
# author: "Daniel Resende"
# date: "January 24, 2015"
# output: html_document
# ---
# Always loads
source('~/workspace/math/constants.R')
cat("\014")
# Chart
pallete <- c('1B305D', '32557B', 'A59750', '6B7D31', '3B461B')
pallete <- paste('#', pallete, sep='')
gray <- '#000000'
# Function
f <- function(x) {
(-4/3*x^2 - 2*x + 5) * sin(x)
}
xDomain <- seq(-4, 4, by=.02)
# yDomain <- range(y)
xRange <- range(xDomain)
# yRange <- yDomain
par(mfrow=c(1,1))
curve(f, xlim=xRange, n=201, main="Tangents: x^3 + x^2 - x")
abline(h = 0, v = 0, col = "gray30")
# png("derivatives.png", width=960, height=590)
x <- sapply(xDomain, function(x, n=.1) {
slope <- (f(x + n) - f(x - n)) / (2*n)
intercept <- f(x) - slope*x
abline(a=intercept, b=slope, col=gray)
})
# dev.off()
| /002-drawing-tangents/main.R | no_license | resendedaniel/math2 | R | false | false | 841 | r | # ---
# title: "1"
# objective: "Function as tangents"
# author: "Daniel Resende"
# date: "January 24, 2015"
# output: html_document
# ---
# Always loads
source('~/workspace/math/constants.R')
cat("\014")
# Chart
pallete <- c('1B305D', '32557B', 'A59750', '6B7D31', '3B461B')
pallete <- paste('#', pallete, sep='')
gray <- '#000000'
# Function
f <- function(x) {
(-4/3*x^2 - 2*x + 5) * sin(x)
}
xDomain <- seq(-4, 4, by=.02)
# yDomain <- range(y)
xRange <- range(xDomain)
# yRange <- yDomain
par(mfrow=c(1,1))
curve(f, xlim=xRange, n=201, main="Tangents: x^3 + x^2 - x")
abline(h = 0, v = 0, col = "gray30")
# png("derivatives.png", width=960, height=590)
x <- sapply(xDomain, function(x, n=.1) {
slope <- (f(x + n) - f(x - n)) / (2*n)
intercept <- f(x) - slope*x
abline(a=intercept, b=slope, col=gray)
})
# dev.off()
|
#source("http://bioconductor.org/biocLite.R")
#biocLite("ShortRead")
library(reshape)
library(stringr)
library(plyr)
library(ShortRead)
require(RMySQL)
session <- dbConnect(MySQL(), host="XXX", db="YYY",user="ZZZ", password="QQQ")
reads=readFastq("reads.out.fq")
filenames=list.files(".",pattern="*.table")
samples=str_extract(filenames, "^(.*?)\\.")
samples=str_replace_all(samples,"R\\d|_|_\\d|\\.","")
classes_files=cbind(samples,filenames)
get_reads = function(row){
sample = row[1]
filename = row[2]
if( file.info(filename)$size < 10 ){
print(paste("! Skipping",sample, filename))
NA
}else{
print(paste(sample, filename))
dat=read.table(filename,as.is=T)
dat=cbind(rep(sample,length(dat$V1)),dat)
names(dat) = c("sample","read_id","reference","identity","score")
dat=data.frame(lapply(dat, as.character), stringsAsFactors=FALSE)
dat$identity=as.numeric(dat$identity)
dat$score=as.numeric(dat$score)
dat
}
}
dd=get_reads(classes_files[1,])
for(i in c(2:(length(classes_files[,1])))){
row=classes_files[i,]
tmp=get_reads(row)
if(is.data.frame(tmp)){
dd=rbind(dd,get_reads(row))
}
}
ids=as.vector(reads@id)
ids=str_replace_all(str_extract(ids, "^(.*?)\\t"), "\\t","")
seqs=as.vector(as.character(reads@sread))
rr=data.frame(read_id=ids,sequence=seqs, stringsAsFactors=FALSE)
res=merge(dd,rr,by=c("read_id"),all=T)
write.table(res,file="merA.csv", quote=T, col.names=T, row.names=F, sep="\t")
# saving the table
#
session <- dbConnect(MySQL(), host="test-mysql.toulouse.inra.fr", db="funnymat",user="funnymat", password="XXX")
#
sql_load_table2=cbind(dd$sample,dd$read_id,dd$reference,dd$identity,dd$score)
write.table(sql_load_table2,file="merA.load2.txt", quote=F, col.names=F, row.names=F, sep="\t")
#LOAD DATA LOCAL INFILE '/work/psenin/fm/KEGG_works/marisol/merB/merB.load2.txt' INTO table merB_hits FIELDS TERMINATED BY '\t' LINES TERMINATED BY '\n';
#
#
sql_load_table1=unique(cbind(dd$reference))
sql_load_table1=cbind(seq(1:length(sql_load_table1)),sql_load_table1)
write.table(sql_load_table1,file="merB.load1.txt", quote=F, col.names=F, row.names=F, sep="\t")
#LOAD DATA LOCAL INFILE '/work/psenin/fm/KEGG_works/marisol/merB/merB.load1.txt' INTO table merB FIELDS TERMINATED BY '\t' LINES TERMINATED BY '\n';
references <- dbGetQuery(session, "select distinct(name) from merA")
samples <- dbGetQuery(session, "select distinct(tag) from merA_hits")
grid <- expand.grid(sample=as.character(unlist(samples)), reference=as.character(unlist(references)))
grid <- cbind(as.character(grid$sample), as.character(grid$reference))
sample_summary = function(X) {
sample = X[1]
reference = X[2]
res = dbGetQuery(session,
paste("select count(*) from merA_hits where tag=",shQuote(sample),
" AND subject_id=",shQuote(reference),";",sep=""))
as.numeric(res)
}
dd=apply(grid,1,sample_summary)
res=data.frame(sample=grid[,1],col=grid[,2],value=as.numeric(dd),stringsAsFactors=F)
ft=cast(res,col~sample)
names(ft)=c("reference",as.character(unlist(samples)))
write.table(ft,file="merA_table.csv", quote=F, col.names=T, row.names=F, sep="\t")
| /RCode/merA.R | no_license | seninp-bioinfo/jKEGG | R | false | false | 3,137 | r | #source("http://bioconductor.org/biocLite.R")
#biocLite("ShortRead")
library(reshape)
library(stringr)
library(plyr)
library(ShortRead)
require(RMySQL)
session <- dbConnect(MySQL(), host="XXX", db="YYY",user="ZZZ", password="QQQ")
reads=readFastq("reads.out.fq")
filenames=list.files(".",pattern="*.table")
samples=str_extract(filenames, "^(.*?)\\.")
samples=str_replace_all(samples,"R\\d|_|_\\d|\\.","")
classes_files=cbind(samples,filenames)
get_reads = function(row){
sample = row[1]
filename = row[2]
if( file.info(filename)$size < 10 ){
print(paste("! Skipping",sample, filename))
NA
}else{
print(paste(sample, filename))
dat=read.table(filename,as.is=T)
dat=cbind(rep(sample,length(dat$V1)),dat)
names(dat) = c("sample","read_id","reference","identity","score")
dat=data.frame(lapply(dat, as.character), stringsAsFactors=FALSE)
dat$identity=as.numeric(dat$identity)
dat$score=as.numeric(dat$score)
dat
}
}
dd=get_reads(classes_files[1,])
for(i in c(2:(length(classes_files[,1])))){
row=classes_files[i,]
tmp=get_reads(row)
if(is.data.frame(tmp)){
dd=rbind(dd,get_reads(row))
}
}
ids=as.vector(reads@id)
ids=str_replace_all(str_extract(ids, "^(.*?)\\t"), "\\t","")
seqs=as.vector(as.character(reads@sread))
rr=data.frame(read_id=ids,sequence=seqs, stringsAsFactors=FALSE)
res=merge(dd,rr,by=c("read_id"),all=T)
write.table(res,file="merA.csv", quote=T, col.names=T, row.names=F, sep="\t")
# saving the table
#
session <- dbConnect(MySQL(), host="test-mysql.toulouse.inra.fr", db="funnymat",user="funnymat", password="XXX")
#
sql_load_table2=cbind(dd$sample,dd$read_id,dd$reference,dd$identity,dd$score)
write.table(sql_load_table2,file="merA.load2.txt", quote=F, col.names=F, row.names=F, sep="\t")
#LOAD DATA LOCAL INFILE '/work/psenin/fm/KEGG_works/marisol/merB/merB.load2.txt' INTO table merB_hits FIELDS TERMINATED BY '\t' LINES TERMINATED BY '\n';
#
#
sql_load_table1=unique(cbind(dd$reference))
sql_load_table1=cbind(seq(1:length(sql_load_table1)),sql_load_table1)
write.table(sql_load_table1,file="merB.load1.txt", quote=F, col.names=F, row.names=F, sep="\t")
#LOAD DATA LOCAL INFILE '/work/psenin/fm/KEGG_works/marisol/merB/merB.load1.txt' INTO table merB FIELDS TERMINATED BY '\t' LINES TERMINATED BY '\n';
references <- dbGetQuery(session, "select distinct(name) from merA")
samples <- dbGetQuery(session, "select distinct(tag) from merA_hits")
grid <- expand.grid(sample=as.character(unlist(samples)), reference=as.character(unlist(references)))
grid <- cbind(as.character(grid$sample), as.character(grid$reference))
sample_summary = function(X) {
sample = X[1]
reference = X[2]
res = dbGetQuery(session,
paste("select count(*) from merA_hits where tag=",shQuote(sample),
" AND subject_id=",shQuote(reference),";",sep=""))
as.numeric(res)
}
dd=apply(grid,1,sample_summary)
res=data.frame(sample=grid[,1],col=grid[,2],value=as.numeric(dd),stringsAsFactors=F)
ft=cast(res,col~sample)
names(ft)=c("reference",as.character(unlist(samples)))
write.table(ft,file="merA_table.csv", quote=F, col.names=T, row.names=F, sep="\t")
|
library(testthat)
library(TimeWindowMaker2)
test_check("TimeWindowMaker2")
| /tests/testthat.R | no_license | itsaquestion/TimeWindowMaker2 | R | false | false | 76 | r | library(testthat)
library(TimeWindowMaker2)
test_check("TimeWindowMaker2")
|
#' Change the absolute path of a file
#'
#' \code{convertPaths} is simply a wrapper around \code{gsub} for changing the
#' first part of a path.
#' \code{convertRasterPaths} is useful for changing the path to a file-backed
#' raster (e.g., after copying the file to a new location).
#'
#' @param x For \code{convertPaths}, a character vector of file paths.
#' For \code{convertRasterPaths}, a disk-backed \code{RasterLayer}
#' object, or a list of such rasters.
#' @param patterns Character vector containing a pattern to match (see \code{?gsub}).
#' @param replacements Character vector of the same length of \code{patterns}
#' containing replacement text (see \code{?gsub}).
#'
#' @author Eliot McIntire and Alex Chubaty
#' @export
#' @rdname convertPaths
#'
#' @examples
#' filenames <- c("/home/user1/Documents/file.txt", "/Users/user1/Documents/file.txt")
#' oldPaths <- dirname(filenames)
#' newPaths <- c("/home/user2/Desktop", "/Users/user2/Desktop")
#' convertPaths(filenames, oldPaths, newPaths)
#'
#' r1 <- raster::raster(system.file("external/test.grd", package = "raster"))
#' r2 <- raster::raster(system.file("external/rlogo.grd", package = "raster"))
#' rasters <- list(r1, r2)
#' oldPaths <- system.file("external", package = "raster")
#' newPaths <- file.path("~/rasters")
#' rasters <- convertRasterPaths(rasters, oldPaths, newPaths)
#' lapply(rasters, raster::filename)
#'
convertPaths <- function(x, patterns, replacements) {
stopifnot(is(x, "character"))
stopifnot(length(patterns) == length(replacements))
patterns <- normPath(patterns)
replacements <- normPath(replacements)
x <- normPath(x)
for (i in seq_along(patterns)) {
x <- gsub(x = x, pattern = patterns[i], replacement = replacements[i])
}
normPath(x)
}
#' @author Eliot McIntire and Alex Chubaty
#' @export
#' @importFrom raster filename raster
#' @rdname convertPaths
convertRasterPaths <- function(x, patterns, replacements) {
if (is.list(x)) {
x <- lapply(x, convertRasterPaths, patterns, replacements)
} else if (!is.null(x)) {
if (is.character(x)) {
if (length(x) > 1) {
x <- lapply(x, convertRasterPaths, patterns, replacements)
} else {
x <- raster(x)
}
}
x@file@name <- convertPaths(filename(x), patterns, replacements)
}
x # handles null case
}
#' Return the filename(s) from a \code{Raster*} object
#'
#' This is mostly just a wrapper around \code{filename} from the \pkg{raster} package, except that
#' instead of returning an empty string for a \code{RasterStack} object, it will return a vector of
#' length >1 for \code{RasterStack}.
#'
#' @param obj A \code{Raster*} object (i.e., \code{RasterLayer}, \code{RasterStack}, \code{RasterBrick})
#' @param allowMultiple Logical. If \code{TRUE}, the default, then all relevant
#' filenames will be returned, i.e., in cases such as \code{.grd} where multiple files
#' are required. If \code{FALSE}, then only the first file will be returned,
#' e.g., \code{filename.grd}, in the case of default Raster format in R.
#'
#' @author Eliot McIntire
#' @export
#' @rdname Filenames
setGeneric("Filenames", function(obj, allowMultiple = TRUE) {
standardGeneric("Filenames")
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "ANY",
definition = function(obj, allowMultiple) {
NULL
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "Raster",
definition = function(obj, allowMultiple = TRUE) {
fn <- filename(obj)
browser(expr = exists("._Filenames_1"))
if (isTRUE(allowMultiple))
if (endsWith(fn, suffix = "grd"))
fn <- c(fn, gsub("grd$", "gri", fn))
normPath(fn)
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "RasterStack",
definition = function(obj, allowMultiple = TRUE) {
fn <- unlist(lapply(seq_along(names(obj)), function(index)
Filenames(obj[[index]], allowMultiple = allowMultiple)))
dups <- duplicated(fn)
if (any(dups)) {
theNames <- names(fn)
fn <- fn[!dups]
names(fn) <- theNames[!dups]
}
return(fn)
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "environment",
definition = function(obj, allowMultiple = TRUE) {
rastersLogical <- isOrHasRaster(obj)
rasterFilename <- NULL
if (any(rastersLogical)) {
rasterNames <- names(rastersLogical)[rastersLogical]
if (!is.null(rasterNames)) {
no <- names(obj);
names(no) <- no;
nestedOnes <- lapply(no, function(rn) grep(paste0("^", rn, "\\."), rasterNames, value = TRUE))
nestedOnes1 <- nestedOnes[sapply(nestedOnes, function(x) length(x) > 0)]
nonNested <- nestedOnes[sapply(nestedOnes, function(x) length(x) == 0)]
nonNestedRasterNames <- rasterNames[rasterNames %in% names(nonNested)]
diskBacked <- sapply(mget(nonNestedRasterNames, envir = obj), fromDisk)
names(rasterNames) <- rasterNames
rasterFilename <- if (sum(diskBacked) > 0) {
lapply(mget(rasterNames[diskBacked], envir = obj), Filenames,
allowMultiple = allowMultiple)
} else {
NULL
}
if (length(nestedOnes1) > 0) {
rasterFilename2 <- sapply(mget(names(nestedOnes1), envir = obj), Filenames,
allowMultiple = allowMultiple)
rasterFilename <- c(rasterFilename, rasterFilename2)
}
}
}
rasterFilenameDups <- lapply(rasterFilename, duplicated)
rasterFilename <- lapply(names(rasterFilenameDups), function(nam) rasterFilename[[nam]][!rasterFilenameDups[[nam]]])
return(rasterFilename)
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "list",
definition = function(obj, allowMultiple = TRUE) {
## convert a list to an environment -- this is to align it with a simList and environment
Filenames(as.environment(obj),
allowMultiple = allowMultiple)
})
| /R/convertPaths.R | no_license | mdsumner/reproducible | R | false | false | 6,066 | r | #' Change the absolute path of a file
#'
#' \code{convertPaths} is simply a wrapper around \code{gsub} for changing the
#' first part of a path.
#' \code{convertRasterPaths} is useful for changing the path to a file-backed
#' raster (e.g., after copying the file to a new location).
#'
#' @param x For \code{convertPaths}, a character vector of file paths.
#' For \code{convertRasterPaths}, a disk-backed \code{RasterLayer}
#' object, or a list of such rasters.
#' @param patterns Character vector containing a pattern to match (see \code{?gsub}).
#' @param replacements Character vector of the same length of \code{patterns}
#' containing replacement text (see \code{?gsub}).
#'
#' @author Eliot McIntire and Alex Chubaty
#' @export
#' @rdname convertPaths
#'
#' @examples
#' filenames <- c("/home/user1/Documents/file.txt", "/Users/user1/Documents/file.txt")
#' oldPaths <- dirname(filenames)
#' newPaths <- c("/home/user2/Desktop", "/Users/user2/Desktop")
#' convertPaths(filenames, oldPaths, newPaths)
#'
#' r1 <- raster::raster(system.file("external/test.grd", package = "raster"))
#' r2 <- raster::raster(system.file("external/rlogo.grd", package = "raster"))
#' rasters <- list(r1, r2)
#' oldPaths <- system.file("external", package = "raster")
#' newPaths <- file.path("~/rasters")
#' rasters <- convertRasterPaths(rasters, oldPaths, newPaths)
#' lapply(rasters, raster::filename)
#'
convertPaths <- function(x, patterns, replacements) {
stopifnot(is(x, "character"))
stopifnot(length(patterns) == length(replacements))
patterns <- normPath(patterns)
replacements <- normPath(replacements)
x <- normPath(x)
for (i in seq_along(patterns)) {
x <- gsub(x = x, pattern = patterns[i], replacement = replacements[i])
}
normPath(x)
}
#' @author Eliot McIntire and Alex Chubaty
#' @export
#' @importFrom raster filename raster
#' @rdname convertPaths
convertRasterPaths <- function(x, patterns, replacements) {
if (is.list(x)) {
x <- lapply(x, convertRasterPaths, patterns, replacements)
} else if (!is.null(x)) {
if (is.character(x)) {
if (length(x) > 1) {
x <- lapply(x, convertRasterPaths, patterns, replacements)
} else {
x <- raster(x)
}
}
x@file@name <- convertPaths(filename(x), patterns, replacements)
}
x # handles null case
}
#' Return the filename(s) from a \code{Raster*} object
#'
#' This is mostly just a wrapper around \code{filename} from the \pkg{raster} package, except that
#' instead of returning an empty string for a \code{RasterStack} object, it will return a vector of
#' length >1 for \code{RasterStack}.
#'
#' @param obj A \code{Raster*} object (i.e., \code{RasterLayer}, \code{RasterStack}, \code{RasterBrick})
#' @param allowMultiple Logical. If \code{TRUE}, the default, then all relevant
#' filenames will be returned, i.e., in cases such as \code{.grd} where multiple files
#' are required. If \code{FALSE}, then only the first file will be returned,
#' e.g., \code{filename.grd}, in the case of default Raster format in R.
#'
#' @author Eliot McIntire
#' @export
#' @rdname Filenames
setGeneric("Filenames", function(obj, allowMultiple = TRUE) {
standardGeneric("Filenames")
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "ANY",
definition = function(obj, allowMultiple) {
NULL
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "Raster",
definition = function(obj, allowMultiple = TRUE) {
fn <- filename(obj)
browser(expr = exists("._Filenames_1"))
if (isTRUE(allowMultiple))
if (endsWith(fn, suffix = "grd"))
fn <- c(fn, gsub("grd$", "gri", fn))
normPath(fn)
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "RasterStack",
definition = function(obj, allowMultiple = TRUE) {
fn <- unlist(lapply(seq_along(names(obj)), function(index)
Filenames(obj[[index]], allowMultiple = allowMultiple)))
dups <- duplicated(fn)
if (any(dups)) {
theNames <- names(fn)
fn <- fn[!dups]
names(fn) <- theNames[!dups]
}
return(fn)
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "environment",
definition = function(obj, allowMultiple = TRUE) {
rastersLogical <- isOrHasRaster(obj)
rasterFilename <- NULL
if (any(rastersLogical)) {
rasterNames <- names(rastersLogical)[rastersLogical]
if (!is.null(rasterNames)) {
no <- names(obj);
names(no) <- no;
nestedOnes <- lapply(no, function(rn) grep(paste0("^", rn, "\\."), rasterNames, value = TRUE))
nestedOnes1 <- nestedOnes[sapply(nestedOnes, function(x) length(x) > 0)]
nonNested <- nestedOnes[sapply(nestedOnes, function(x) length(x) == 0)]
nonNestedRasterNames <- rasterNames[rasterNames %in% names(nonNested)]
diskBacked <- sapply(mget(nonNestedRasterNames, envir = obj), fromDisk)
names(rasterNames) <- rasterNames
rasterFilename <- if (sum(diskBacked) > 0) {
lapply(mget(rasterNames[diskBacked], envir = obj), Filenames,
allowMultiple = allowMultiple)
} else {
NULL
}
if (length(nestedOnes1) > 0) {
rasterFilename2 <- sapply(mget(names(nestedOnes1), envir = obj), Filenames,
allowMultiple = allowMultiple)
rasterFilename <- c(rasterFilename, rasterFilename2)
}
}
}
rasterFilenameDups <- lapply(rasterFilename, duplicated)
rasterFilename <- lapply(names(rasterFilenameDups), function(nam) rasterFilename[[nam]][!rasterFilenameDups[[nam]]])
return(rasterFilename)
})
#' @export
#' @rdname Filenames
setMethod(
"Filenames",
signature = "list",
definition = function(obj, allowMultiple = TRUE) {
## convert a list to an environment -- this is to align it with a simList and environment
Filenames(as.environment(obj),
allowMultiple = allowMultiple)
})
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/bar.R
\name{bar}
\alias{bar}
\title{Bar Plot data file}
\usage{
bar(data, map, value = c("count", "identity"))
}
\arguments{
\item{data}{data.frame to make plot-ready data for}
\item{map}{data.frame with at least two columns (id, plotRef) indicating a variable sourceId and its position in the plot}
\item{value}{String indicating how to calculate y-values ('identity', 'count')}
}
\value{
character name of json file containing plot-ready data
}
\description{
This function returns the name of a json file containing
plot-ready data with one row per group (per panel). Columns
'label' and 'value' contain the raw data for plotting. Column
'group' and 'panel' specify the group the series data belongs to.
There are two options to calculate y-values for plotting.
1) raw 'identity' of values from data.table input
2) 'count' occurances of values from data.table input
}
| /man/bar.Rd | no_license | d-callan/plot.data | R | false | true | 953 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/bar.R
\name{bar}
\alias{bar}
\title{Bar Plot data file}
\usage{
bar(data, map, value = c("count", "identity"))
}
\arguments{
\item{data}{data.frame to make plot-ready data for}
\item{map}{data.frame with at least two columns (id, plotRef) indicating a variable sourceId and its position in the plot}
\item{value}{String indicating how to calculate y-values ('identity', 'count')}
}
\value{
character name of json file containing plot-ready data
}
\description{
This function returns the name of a json file containing
plot-ready data with one row per group (per panel). Columns
'label' and 'value' contain the raw data for plotting. Column
'group' and 'panel' specify the group the series data belongs to.
There are two options to calculate y-values for plotting.
1) raw 'identity' of values from data.table input
2) 'count' occurances of values from data.table input
}
|
# Master File
library(dplyr)
library(ggplot2)
library(shiny)
library(leaflet)
library(mapview)
library(rsconnect)
library(geojsonio)
library(htmltools)
library(reshape2)
source("code.R")
source("EmanHardcode.R")
# Define server logic required to draw a histogram
shinyServer(function(input, output) {
# Home Tab
# Map Tab
output$mymap <- renderLeaflet({
#You input a state and you get the percentage as an output for speeding, not distracted, alcohol impaired, and not involved in
if(input$states == "Speeding") {
states = full_data$Percentage.Of.Drivers.Involved.In.Fatal.Collisions.Who.Were.Speeding
} else if(input$states == "Not distracted") {
states = full_data$Percentage.Of.Drivers.Involved.In.Fatal.Collisions.Who.Were.Not.Distracted
} else if(input$states == "Alcohol impaired") {
states = full_data$Percentage.Of.Drivers.Involved.In.Fatal.Collisions.Who.Were.Alcohol.Impaired
} else {
states = full_data$Percentage.Of.Drivers.Involved.In.Fatal.Collisions.Who.Had.Not.Been.Involved.In.Any.Previous.Accidents
}
#For the display of the markers
string1 <- paste(strong(span("State:", style = "color:#0083C4")), full_data$State) #String of the state name, "State:" is bold and colored to match the color of the marker
string2 <- paste("The percentage is:", states, "%") #String of the percentage of the options for that specific state
popup <- paste(sep = "<br/>", string1, string2) #Line break of both strings
leaflet(state_data) %>%
setView(-96, 37.8, 4) %>% #Lat and long coordinates for the map boundaries
addProviderTiles("MapBox", options = providerTileOptions(
id = "mapbox.light", #This creates an outline of every state in the United States
accessToken = Sys.getenv('pk.eyJ1IjoiaHdhaGVlZWQiLCJhIjoiY2phc21jYnY1NHNibjJxcGxseG9vMzl4cSJ9.xTnmU0DdOSrPePsTnlRdgg'))) %>%
addTiles() %>% #Adds the background of all continents, makes it look like an actual map rather than just the outline
addPolygons(weight = 2,
opacity = 1,
color = "black", #Outline adjustments
dashArray = "1",
fillOpacity = 0.2) %>%
addMarkers(data = full_data, lng = full_data$lng, lat = full_data$lat, popup=popup) %>% #The markers placed on the map
setView(lng = -95.85, lat = 38.75, zoom = 5) #The boundary that is first displayed when opened
})
# Car Insurance Tab
output$Information <- renderPlot({
State_Names <- rep(c("Alabama", "California", "Delaware", "Florida", "Georgia", "Hawaii", "Idaho", "Kansas", "Louisiana", "Maine", "Nebraska", "Ohio",
"Pennsylvania", "Tennessee", "Utah", "Vermont", "Washington"), 2)
Car_Insurance_and_Losses <- rep(c(Specific_States_melt$variable))
Colors <- c(rep("Car Insurance Premiums", 1), rep("Losses incurred by insurance companies for collisions per insured driver", 1))
Insurance_Losses_Data <- data.frame( # create dataframe to select variable states, and costs for ggplot
Variable <- factor(c(Car_Insurance_and_Losses)),
States <- factor(c(State_Names)),
costs <- c(Specific_States_melt$value)
)
data <- filter(Insurance_Losses_Data, States == input$State) # filters the specific state in the datatable
ggplot(data, aes(x = data$State, y = data$costs, fill = Colors)) + # creates ggplot
geom_bar(colour = "Black", stat = "identity",
position=position_dodge(),
size=.3) + # Thinner lines
xlab("Name of State") + ylab("Premium and Losses costs") + # labels the x and y axis
ggtitle("Car Insurance Premiums vs Losses incurred by insurance companies for collisions per insured driver") + # Set title
theme_dark() # gives dark background
})
# Comparision Tab
dataPlot <- read.csv(file = 'data/bad-drivers.csv', stringsAsFactors = FALSE)
# Change column names of data frame
colnames(dataPlot) <- c("State", "Number.Collisions", "Collisions.Speeding", "Collisions.Alcohol", "Collisions.Not.Distracted", "Collisions.No.Previous", "Car.Insurance", "Losses.By.Insurance")
# Compute US average of fatal collisions per billion miles
usMeanDataFrame <- summarize(dataPlot, mean = mean(Number.Collisions))
usMean <- usMeanDataFrame[1,1]
output$higherAverage <- renderPlot({
# filter out data frame for states higher than US mean
higherStates <- filter(dataPlot, Number.Collisions > usMean)
# draw the histogram with the specified number of bins
ggplot(higherStates) + geom_point(aes_string(x = "State", y = input$higherY, color = input$higherColorType)) + ggtitle("Analysis of States Data Higher than US Mean") + theme(axis.text.x=element_text(angle=90,hjust=1))
})
output$lowerAverage <- renderPlot({
# filter out data frame for states lower than US mean
lowerStates <- filter(dataPlot, Number.Collisions < usMean)
# draw the histogram with the specified number of bins
ggplot(lowerStates) + geom_point(aes_string(x = "State", y = input$lowerY, color = input$lowerColorType)) + ggtitle("Analysis of States Data Lower than US Mean") + theme(axis.text.x=element_text(angle=90,hjust=1))
})
})
| /server.R | no_license | nehay100/SHENFinalProject | R | false | false | 5,450 | r | # Master File
library(dplyr)
library(ggplot2)
library(shiny)
library(leaflet)
library(mapview)
library(rsconnect)
library(geojsonio)
library(htmltools)
library(reshape2)
source("code.R")
source("EmanHardcode.R")
# Define server logic required to draw a histogram
shinyServer(function(input, output) {
# Home Tab
# Map Tab
output$mymap <- renderLeaflet({
#You input a state and you get the percentage as an output for speeding, not distracted, alcohol impaired, and not involved in
if(input$states == "Speeding") {
states = full_data$Percentage.Of.Drivers.Involved.In.Fatal.Collisions.Who.Were.Speeding
} else if(input$states == "Not distracted") {
states = full_data$Percentage.Of.Drivers.Involved.In.Fatal.Collisions.Who.Were.Not.Distracted
} else if(input$states == "Alcohol impaired") {
states = full_data$Percentage.Of.Drivers.Involved.In.Fatal.Collisions.Who.Were.Alcohol.Impaired
} else {
states = full_data$Percentage.Of.Drivers.Involved.In.Fatal.Collisions.Who.Had.Not.Been.Involved.In.Any.Previous.Accidents
}
#For the display of the markers
string1 <- paste(strong(span("State:", style = "color:#0083C4")), full_data$State) #String of the state name, "State:" is bold and colored to match the color of the marker
string2 <- paste("The percentage is:", states, "%") #String of the percentage of the options for that specific state
popup <- paste(sep = "<br/>", string1, string2) #Line break of both strings
leaflet(state_data) %>%
setView(-96, 37.8, 4) %>% #Lat and long coordinates for the map boundaries
addProviderTiles("MapBox", options = providerTileOptions(
id = "mapbox.light", #This creates an outline of every state in the United States
accessToken = Sys.getenv('pk.eyJ1IjoiaHdhaGVlZWQiLCJhIjoiY2phc21jYnY1NHNibjJxcGxseG9vMzl4cSJ9.xTnmU0DdOSrPePsTnlRdgg'))) %>%
addTiles() %>% #Adds the background of all continents, makes it look like an actual map rather than just the outline
addPolygons(weight = 2,
opacity = 1,
color = "black", #Outline adjustments
dashArray = "1",
fillOpacity = 0.2) %>%
addMarkers(data = full_data, lng = full_data$lng, lat = full_data$lat, popup=popup) %>% #The markers placed on the map
setView(lng = -95.85, lat = 38.75, zoom = 5) #The boundary that is first displayed when opened
})
# Car Insurance Tab
output$Information <- renderPlot({
State_Names <- rep(c("Alabama", "California", "Delaware", "Florida", "Georgia", "Hawaii", "Idaho", "Kansas", "Louisiana", "Maine", "Nebraska", "Ohio",
"Pennsylvania", "Tennessee", "Utah", "Vermont", "Washington"), 2)
Car_Insurance_and_Losses <- rep(c(Specific_States_melt$variable))
Colors <- c(rep("Car Insurance Premiums", 1), rep("Losses incurred by insurance companies for collisions per insured driver", 1))
Insurance_Losses_Data <- data.frame( # create dataframe to select variable states, and costs for ggplot
Variable <- factor(c(Car_Insurance_and_Losses)),
States <- factor(c(State_Names)),
costs <- c(Specific_States_melt$value)
)
data <- filter(Insurance_Losses_Data, States == input$State) # filters the specific state in the datatable
ggplot(data, aes(x = data$State, y = data$costs, fill = Colors)) + # creates ggplot
geom_bar(colour = "Black", stat = "identity",
position=position_dodge(),
size=.3) + # Thinner lines
xlab("Name of State") + ylab("Premium and Losses costs") + # labels the x and y axis
ggtitle("Car Insurance Premiums vs Losses incurred by insurance companies for collisions per insured driver") + # Set title
theme_dark() # gives dark background
})
# Comparision Tab
dataPlot <- read.csv(file = 'data/bad-drivers.csv', stringsAsFactors = FALSE)
# Change column names of data frame
colnames(dataPlot) <- c("State", "Number.Collisions", "Collisions.Speeding", "Collisions.Alcohol", "Collisions.Not.Distracted", "Collisions.No.Previous", "Car.Insurance", "Losses.By.Insurance")
# Compute US average of fatal collisions per billion miles
usMeanDataFrame <- summarize(dataPlot, mean = mean(Number.Collisions))
usMean <- usMeanDataFrame[1,1]
output$higherAverage <- renderPlot({
# filter out data frame for states higher than US mean
higherStates <- filter(dataPlot, Number.Collisions > usMean)
# draw the histogram with the specified number of bins
ggplot(higherStates) + geom_point(aes_string(x = "State", y = input$higherY, color = input$higherColorType)) + ggtitle("Analysis of States Data Higher than US Mean") + theme(axis.text.x=element_text(angle=90,hjust=1))
})
output$lowerAverage <- renderPlot({
# filter out data frame for states lower than US mean
lowerStates <- filter(dataPlot, Number.Collisions < usMean)
# draw the histogram with the specified number of bins
ggplot(lowerStates) + geom_point(aes_string(x = "State", y = input$lowerY, color = input$lowerColorType)) + ggtitle("Analysis of States Data Lower than US Mean") + theme(axis.text.x=element_text(angle=90,hjust=1))
})
})
|
# Demonstrating side effects
x<-1
good <- function() { x <- 5}
good()
print(x)
## [1] 1
bad <- function() { x <<- 5}
bad()
print(x)
## [1] 5
| /datascience_course/δΈθͺ²η¨εΌη’Ό/week_2/sideEff.R | no_license | bill0812/course_study | R | false | false | 141 | r | # Demonstrating side effects
x<-1
good <- function() { x <- 5}
good()
print(x)
## [1] 1
bad <- function() { x <<- 5}
bad()
print(x)
## [1] 5
|
###
### imomknown.R.R
###
imomknown <- function(theta1hat,V1,n,nuisance.theta,g=1,nu=1,theta0,sigma,method='adapt',B=10^5) {
if (missing(sigma)) stop('sigma must be specified')
if (missing(theta0)) theta0 <- rep(0,length(theta1hat))
f <- function(z) { ans <- (n*gi/z)^((nu+p1)/2) * exp(-n*gi/z); ans[z==0] <- 0; return(ans) }
p1 <- length(theta1hat)
l <- theta1hat-theta0; l <- matrix(l,nrow=1) %*% solve(V1) %*% matrix(l,ncol=1) / sigma^2 #noncentr param
m <- double(length(g))
if (method=='MC') {
z <- rchisq(B,df=p1,ncp=l)
for (i in 1:length(m)) { gi <- g[i]; m[i] <- mean(f(z)) }
} else if (method=='adapt') {
f2 <- function(z) { return(f(z)*dchisq(z,df=p1,ncp=l)) }
for (i in 1:length(m)) { gi <- g[i]; m[i] <- integrate(f2,0,Inf)$value }
} else {
stop('method must be adapt or MC')
}
bf <- exp((p1/2)*log(2/(n*g)) + lgamma(p1/2)-lgamma(nu/2) + .5*l) * m
return(bf)
}
| /mombf/R/imomknown.R | no_license | ingted/R-Examples | R | false | false | 886 | r | ###
### imomknown.R.R
###
imomknown <- function(theta1hat,V1,n,nuisance.theta,g=1,nu=1,theta0,sigma,method='adapt',B=10^5) {
if (missing(sigma)) stop('sigma must be specified')
if (missing(theta0)) theta0 <- rep(0,length(theta1hat))
f <- function(z) { ans <- (n*gi/z)^((nu+p1)/2) * exp(-n*gi/z); ans[z==0] <- 0; return(ans) }
p1 <- length(theta1hat)
l <- theta1hat-theta0; l <- matrix(l,nrow=1) %*% solve(V1) %*% matrix(l,ncol=1) / sigma^2 #noncentr param
m <- double(length(g))
if (method=='MC') {
z <- rchisq(B,df=p1,ncp=l)
for (i in 1:length(m)) { gi <- g[i]; m[i] <- mean(f(z)) }
} else if (method=='adapt') {
f2 <- function(z) { return(f(z)*dchisq(z,df=p1,ncp=l)) }
for (i in 1:length(m)) { gi <- g[i]; m[i] <- integrate(f2,0,Inf)$value }
} else {
stop('method must be adapt or MC')
}
bf <- exp((p1/2)*log(2/(n*g)) + lgamma(p1/2)-lgamma(nu/2) + .5*l) * m
return(bf)
}
|
`mu.varknown` <-
function(len,lambda,n0,level=0.95)
{
# Returns the minimal sample size to ensure an average prob. coverage of
# 'level' with a symmetric region about the posterior mean
# when sampling from a normal distribution with known variance (Adcock, 1988)
# Using the same notation as Adcock
var <- 1/lambda
d <- len/2
m <- n0
m <- ceiling(((qnorm((level+1)/2)/d)^2)*var-m)
max(0, m)
}
| /R/mu.varknown.R | no_license | cran/SampleSizeMeans | R | false | false | 435 | r | `mu.varknown` <-
function(len,lambda,n0,level=0.95)
{
# Returns the minimal sample size to ensure an average prob. coverage of
# 'level' with a symmetric region about the posterior mean
# when sampling from a normal distribution with known variance (Adcock, 1988)
# Using the same notation as Adcock
var <- 1/lambda
d <- len/2
m <- n0
m <- ceiling(((qnorm((level+1)/2)/d)^2)*var-m)
max(0, m)
}
|
## timestart timediff 2020-06-20
#' @title Collect the difftime between two events
#'
#' @description
#' \code{timestart} starts the timer and saved the value in an object named
#' \code{time0} stored in \code{.GlobalEnv}.
#'
#' \code{timediff} stops the timer, remove the \code{time0} objet from \code{.GlobalEnv}
#' and prints the duration in seconds between the two events.
#'
#' \code{timestart} and \code{timediff} are fully independant from the R6 class
#' \code{timeR} and the objects \code{createTimer} or \code{getTimer}. They use
#' \code{\link{proc.time}} instead.
#'
#' @return
#' A single numeric value that represents a duration in seconds.
#'
#' @examples
#' timestart()
#' Sys.sleep(2)
#' timediff()
#'
#' @export
#' @name timestart
timestart <- function() {
gc(FALSE)
time0 <- proc.time()["elapsed"]
time0 <<- time0
}
#' @export
#' @rdname timestart
timediff <- function() {
t1 <- proc.time()["elapsed"]
t0 <- get("time0", envir = .GlobalEnv)
# remove("time0", envir = .GlobalEnv)
time0 <- NA
time0 <<- time0
unname(t1 - t0)
}
| /R/timestart-timediff.R | no_license | cran/NNbenchmark | R | false | false | 1,137 | r | ## timestart timediff 2020-06-20
#' @title Collect the difftime between two events
#'
#' @description
#' \code{timestart} starts the timer and saved the value in an object named
#' \code{time0} stored in \code{.GlobalEnv}.
#'
#' \code{timediff} stops the timer, remove the \code{time0} objet from \code{.GlobalEnv}
#' and prints the duration in seconds between the two events.
#'
#' \code{timestart} and \code{timediff} are fully independant from the R6 class
#' \code{timeR} and the objects \code{createTimer} or \code{getTimer}. They use
#' \code{\link{proc.time}} instead.
#'
#' @return
#' A single numeric value that represents a duration in seconds.
#'
#' @examples
#' timestart()
#' Sys.sleep(2)
#' timediff()
#'
#' @export
#' @name timestart
timestart <- function() {
gc(FALSE)
time0 <- proc.time()["elapsed"]
time0 <<- time0
}
#' @export
#' @rdname timestart
timediff <- function() {
t1 <- proc.time()["elapsed"]
t0 <- get("time0", envir = .GlobalEnv)
# remove("time0", envir = .GlobalEnv)
time0 <- NA
time0 <<- time0
unname(t1 - t0)
}
|
# Coded by Anderson Borba data: 22/06/2020 version 1.0
# The total log-likelihood presented in equation (X) in the article
# Article to appear
# XXXXX
# Anderson A. de Borba, MaurΔ±Μcio Marengoni, and Alejandro C Frery
#
# - The program reads an image of two halves 400 X 400 (channels hh, hv and vv) finds the parameters (rho)
# by the BFGS method, and uses in function l(j) , calculating the point of max/min (edge evidence) by
# the GenSA method.
# - The output of the program is a 400 size vector with the edge evidence for each line
# of the recorded image in a *.txt file
# obs: 1) Change the channels in the input and output files.
# 2) The code make to run samples in two halves with proposed sigmas in \cite{gamf} (see article).
# 3) Disable the print in file after running the tests of interest in order not to modify files unduly.
rm(list = ls())
require(ggplot2)
require(latex2exp)
require(GenSA)
require(maxLik)
#
source("func_obj_l_prod_mag.r")
source("loglike_prod_mag.r")
source("loglikd_prod_mag.r")
source("estima_L.r")
# Programa principal
setwd("../..")
setwd("Data")
# canais hh, hv, and vv
# canais para a + bi
mat1 <- scan('Phantom_gamf_0.000_1_2_1.txt')
mat2 <- scan('Phantom_gamf_0.000_1_2_2.txt')
mat3 <- scan('Phantom_gamf_0.000_1_2_4.txt')
mat4 <- scan('Phantom_gamf_0.000_1_2_5.txt')
setwd("..")
setwd("Code/Code_art_rem_sen_2020")
mat1 <- matrix(mat1, ncol = 400, byrow = TRUE)
mat2 <- matrix(mat2, ncol = 400, byrow = TRUE)
mat3 <- matrix(mat3, ncol = 400, byrow = TRUE)
mat4 <- matrix(mat4, ncol = 400, byrow = TRUE)
d <- dim(mat1)
nrows <- d[1]
ncols <- d[2]
# Loop para toda a imagem
evidencias <- rep(0, nrows)
evidencias_valores <- rep(0, nrows)
xev <- seq(1, nrows, 1 )
#L <- 4
#for (k in 1 : nrows){
for (k in 120 : 120){
print(k)
N <- ncols
z1 <- rep(0, N)
z2 <- rep(0, N)
z3 <- rep(0, N)
z4 <- rep(0, N)
zaux1 <- rep(0, N)
z1 <- mat1[k,1:N]
z2 <- mat2[k,1:N]
z3 <- mat3[k,1:N]
z4 <- mat4[k,1:N]
conta = 0
for (i in 1 : N){
if (z1[i] > 0 && z2[i] > 0){
conta <- conta + 1
zaux1[conta] <- (z3[i]^2 + z4[i]^2)^0.5 / (z1[i] * z2[i])^0.5
}
}
indx <- which(zaux1 != 0)
N <- length(indx)
z <- rep(0, N)
z[1: N] <- zaux1[1: N]
matdf1 <- matrix(0, nrow = N, ncol = 1)
matdf2 <- matrix(0, nrow = N, ncol = 1)
L <- 4
for (j in 1 : (N - 1)){
r1 <- 0.01
res1 <- maxBFGS(loglike_prod_mag, start=c(r1))
matdf1[j, 1] <- res1$estimate[1]
r1 <- 0.01
res2 <- maxBFGS(loglikd_prod_mag, start=c(r1))
matdf2[j, 1] <- res2$estimate[1]
}
cf <- 14
lower <- as.numeric(cf)
upper <- as.numeric(N - cf)
out <- GenSA(lower = lower, upper = upper, fn = func_obj_l_prod_mag, control=list(maxit = 100))
evidencias[k] <- out$par
print(evidencias[k])
evidencias_valores[k] <- out$value
}
x <- seq(N - 1)
lobj <- rep(0, N - 1)
for (j in 1 : (N - 1)){
lobj[j] <- func_obj_l_prod_mag(j)
}
df <- data.frame(x, lobj)
p <- ggplot(df, aes(x = x, y = lobj, color = 'darkred')) + geom_line() + xlab(TeX('Pixel $j$')) + ylab(TeX('$l(j)$')) + guides(color=guide_legend(title=NULL)) + scale_color_discrete(labels= lapply(sprintf('$\\sigma_{hh} = %2.0f$', NULL), TeX))
print(p)
# imprime em arquivo no diretorio ~/Data/
#dfev <- data.frame(xev, evidencias)
#names(dfev) <- NULL
#setwd("../..")
#setwd("Data")
#sink("evid_real_flevoland_produto_mag_param_L_rho_1_3.txt")
#print(dfev)
#sink()
#setwd("..")
#setwd("Code/Code_art_rem_sen_2020")
| /Code/Code_art_rem_sen_2020/evidencias_im_sim_param_prod_mag.R | no_license | c-l-k/ufal_mack | R | false | false | 3,476 | r | # Coded by Anderson Borba data: 22/06/2020 version 1.0
# The total log-likelihood presented in equation (X) in the article
# Article to appear
# XXXXX
# Anderson A. de Borba, MaurΔ±Μcio Marengoni, and Alejandro C Frery
#
# - The program reads an image of two halves 400 X 400 (channels hh, hv and vv) finds the parameters (rho)
# by the BFGS method, and uses in function l(j) , calculating the point of max/min (edge evidence) by
# the GenSA method.
# - The output of the program is a 400 size vector with the edge evidence for each line
# of the recorded image in a *.txt file
# obs: 1) Change the channels in the input and output files.
# 2) The code make to run samples in two halves with proposed sigmas in \cite{gamf} (see article).
# 3) Disable the print in file after running the tests of interest in order not to modify files unduly.
rm(list = ls())
require(ggplot2)
require(latex2exp)
require(GenSA)
require(maxLik)
#
source("func_obj_l_prod_mag.r")
source("loglike_prod_mag.r")
source("loglikd_prod_mag.r")
source("estima_L.r")
# Programa principal
setwd("../..")
setwd("Data")
# canais hh, hv, and vv
# canais para a + bi
mat1 <- scan('Phantom_gamf_0.000_1_2_1.txt')
mat2 <- scan('Phantom_gamf_0.000_1_2_2.txt')
mat3 <- scan('Phantom_gamf_0.000_1_2_4.txt')
mat4 <- scan('Phantom_gamf_0.000_1_2_5.txt')
setwd("..")
setwd("Code/Code_art_rem_sen_2020")
mat1 <- matrix(mat1, ncol = 400, byrow = TRUE)
mat2 <- matrix(mat2, ncol = 400, byrow = TRUE)
mat3 <- matrix(mat3, ncol = 400, byrow = TRUE)
mat4 <- matrix(mat4, ncol = 400, byrow = TRUE)
d <- dim(mat1)
nrows <- d[1]
ncols <- d[2]
# Loop para toda a imagem
evidencias <- rep(0, nrows)
evidencias_valores <- rep(0, nrows)
xev <- seq(1, nrows, 1 )
#L <- 4
#for (k in 1 : nrows){
for (k in 120 : 120){
print(k)
N <- ncols
z1 <- rep(0, N)
z2 <- rep(0, N)
z3 <- rep(0, N)
z4 <- rep(0, N)
zaux1 <- rep(0, N)
z1 <- mat1[k,1:N]
z2 <- mat2[k,1:N]
z3 <- mat3[k,1:N]
z4 <- mat4[k,1:N]
conta = 0
for (i in 1 : N){
if (z1[i] > 0 && z2[i] > 0){
conta <- conta + 1
zaux1[conta] <- (z3[i]^2 + z4[i]^2)^0.5 / (z1[i] * z2[i])^0.5
}
}
indx <- which(zaux1 != 0)
N <- length(indx)
z <- rep(0, N)
z[1: N] <- zaux1[1: N]
matdf1 <- matrix(0, nrow = N, ncol = 1)
matdf2 <- matrix(0, nrow = N, ncol = 1)
L <- 4
for (j in 1 : (N - 1)){
r1 <- 0.01
res1 <- maxBFGS(loglike_prod_mag, start=c(r1))
matdf1[j, 1] <- res1$estimate[1]
r1 <- 0.01
res2 <- maxBFGS(loglikd_prod_mag, start=c(r1))
matdf2[j, 1] <- res2$estimate[1]
}
cf <- 14
lower <- as.numeric(cf)
upper <- as.numeric(N - cf)
out <- GenSA(lower = lower, upper = upper, fn = func_obj_l_prod_mag, control=list(maxit = 100))
evidencias[k] <- out$par
print(evidencias[k])
evidencias_valores[k] <- out$value
}
x <- seq(N - 1)
lobj <- rep(0, N - 1)
for (j in 1 : (N - 1)){
lobj[j] <- func_obj_l_prod_mag(j)
}
df <- data.frame(x, lobj)
p <- ggplot(df, aes(x = x, y = lobj, color = 'darkred')) + geom_line() + xlab(TeX('Pixel $j$')) + ylab(TeX('$l(j)$')) + guides(color=guide_legend(title=NULL)) + scale_color_discrete(labels= lapply(sprintf('$\\sigma_{hh} = %2.0f$', NULL), TeX))
print(p)
# imprime em arquivo no diretorio ~/Data/
#dfev <- data.frame(xev, evidencias)
#names(dfev) <- NULL
#setwd("../..")
#setwd("Data")
#sink("evid_real_flevoland_produto_mag_param_L_rho_1_3.txt")
#print(dfev)
#sink()
#setwd("..")
#setwd("Code/Code_art_rem_sen_2020")
|
library(openxlsx) # This library is used to open xlsx files directly in R
library(chron) # This library is used for time manipulation
library(data.table) # Used to join rows one after another
# file = "Tata.xlsx"
# SUS_NO = "9058786086"
# imp_row = 4
# raw_sheet = 1
infile = readline(prompt="Input file: ")
SUS_NO = readline(prompt="Suspect number: ")
imp_row = as.numeric(readline(prompt="Row no. containing headers: "))
raw_sheet = as.numeric(readline(prompt="Raw CDR Sheet number: "))
# Read particular sheet from excel workbook; adjust startRow as per the number of useless rows in the beginnig, remember to remove empty rows
df = read.xlsx(xlsxFile=infile, sheet=raw_sheet, startRow=imp_row, colNames=TRUE, detectDates=TRUE, skipEmptyRows=FALSE)
df <- df[!is.na(df$DURATION),] # remove useless rows from bottom
names(df)[names(df) == 'Calling.No'] <- 'Calling_No' # Do this in case you want to change the header name of a particular column
#Do this in case you want to change the names of all columns (more preferable)
colnames(df) <- c("Calling_No", "Called_No", "Date", "Time", "Duration", "Cell_1", "Cell_2", "Communication_Type",
"IMEI", "IMSI", "Type", "SMSC", "Roam", "Switch", "LRN")
drops <- c("Type","SMSC","Switch","LRN") # These are the columns you want that may be removed, change them as per your will
df = df[ , !(names(df) %in% drops)] # drop the columns from the dataframe
# Make hyphens empty in Cell_1 and Cell_2 columns
df$Cell_1[df$Cell_1 == "-"] <- ""
df$Cell_2[df$Cell_2 == "-"] <- ""
df$Roam[df$Roam == "-"] <- ""
df$Time <- times(as.numeric(df$Time)) # Change time from default decimal format to proper HH::MM::SS format
# Microsoft Excel date gives some offset, remove it using this and change format to proper DD/MM/YY format
df$Date <- format(as.Date(df$Date), "%d/%m/%y")
########################################################################################################################################################
# Create a new dataframe for pivoting according to B_party
# This function has two parameters -> x: phone numbers; t: communication type, supplied when calling this function
f <- function(x, t)
{
return (sum( df$Communication_Type[df$Calling_No == x | df$Called_No == x] == t ))
}
J = rbind(as.matrix(df$Calling_No), as.matrix(df$Called_No)) # all numbers that communicated with our suspect
U = as.matrix(unique(J[J != SUS_NO])) # the same in matrix format, to be used in apply function
rm(J) # delete temporary J
# Gets count of each communication type; MARGIN=1, function is applied for each row of matrix
# x1 means first column of X=U
SMS_IN_List = apply(X=U, MARGIN=1, FUN=function(x1) f(x = x1, t="SMT"))
SMS_OUT_List = apply(X=U, MARGIN=1, FUN=function(x1) f(x = x1, t="SMO"))
CALL_IN_List = apply(X=U, MARGIN=1, FUN=function(x1) f(x = x1, t="MTC"))
CALL_OUT_List = apply(X=U, MARGIN=1, FUN=function(x1) f(x = x1, t="MOC"))
# Total communication count between all B-parties
TOTAL_List = SMS_IN_List+SMS_OUT_List+CALL_IN_List+CALL_OUT_List
# Create pivot dataframe: B-Party and communication count details
df_pivot <- data.frame(U,CALL_IN_List,CALL_OUT_List,SMS_IN_List,SMS_OUT_List,TOTAL_List)
colnames(df_pivot) <- c("B_Party", "IN", "OUT", "SMS_IN", "SMS_OUT", "Total") # give column names
df_pivot <- df_pivot[order(-df_pivot$Total),] # Note the (-) sign: sort in descending order of total communication
# Output pivot details to a file
outfile = readline(prompt="Output file: ")
sheet = readline(prompt="Sheet name: ")
hs <- createStyle(textDecoration = "BOLD", fontColour = "#FFFFFF", fontSize=12, fontName="Arial Narrow", fgFill = "#4F80BD")
write.xlsx(x=df_pivot, file=outfile, sheetName = sheet, headerStyle=hs) | /Tata.R | no_license | aakanksha287/Cell-Phone-Theft-problem | R | false | false | 3,793 | r | library(openxlsx) # This library is used to open xlsx files directly in R
library(chron) # This library is used for time manipulation
library(data.table) # Used to join rows one after another
# file = "Tata.xlsx"
# SUS_NO = "9058786086"
# imp_row = 4
# raw_sheet = 1
infile = readline(prompt="Input file: ")
SUS_NO = readline(prompt="Suspect number: ")
imp_row = as.numeric(readline(prompt="Row no. containing headers: "))
raw_sheet = as.numeric(readline(prompt="Raw CDR Sheet number: "))
# Read particular sheet from excel workbook; adjust startRow as per the number of useless rows in the beginnig, remember to remove empty rows
df = read.xlsx(xlsxFile=infile, sheet=raw_sheet, startRow=imp_row, colNames=TRUE, detectDates=TRUE, skipEmptyRows=FALSE)
df <- df[!is.na(df$DURATION),] # remove useless rows from bottom
names(df)[names(df) == 'Calling.No'] <- 'Calling_No' # Do this in case you want to change the header name of a particular column
#Do this in case you want to change the names of all columns (more preferable)
colnames(df) <- c("Calling_No", "Called_No", "Date", "Time", "Duration", "Cell_1", "Cell_2", "Communication_Type",
"IMEI", "IMSI", "Type", "SMSC", "Roam", "Switch", "LRN")
drops <- c("Type","SMSC","Switch","LRN") # These are the columns you want that may be removed, change them as per your will
df = df[ , !(names(df) %in% drops)] # drop the columns from the dataframe
# Make hyphens empty in Cell_1 and Cell_2 columns
df$Cell_1[df$Cell_1 == "-"] <- ""
df$Cell_2[df$Cell_2 == "-"] <- ""
df$Roam[df$Roam == "-"] <- ""
df$Time <- times(as.numeric(df$Time)) # Change time from default decimal format to proper HH::MM::SS format
# Microsoft Excel date gives some offset, remove it using this and change format to proper DD/MM/YY format
df$Date <- format(as.Date(df$Date), "%d/%m/%y")
########################################################################################################################################################
# Create a new dataframe for pivoting according to B_party
# This function has two parameters -> x: phone numbers; t: communication type, supplied when calling this function
f <- function(x, t)
{
return (sum( df$Communication_Type[df$Calling_No == x | df$Called_No == x] == t ))
}
J = rbind(as.matrix(df$Calling_No), as.matrix(df$Called_No)) # all numbers that communicated with our suspect
U = as.matrix(unique(J[J != SUS_NO])) # the same in matrix format, to be used in apply function
rm(J) # delete temporary J
# Gets count of each communication type; MARGIN=1, function is applied for each row of matrix
# x1 means first column of X=U
SMS_IN_List = apply(X=U, MARGIN=1, FUN=function(x1) f(x = x1, t="SMT"))
SMS_OUT_List = apply(X=U, MARGIN=1, FUN=function(x1) f(x = x1, t="SMO"))
CALL_IN_List = apply(X=U, MARGIN=1, FUN=function(x1) f(x = x1, t="MTC"))
CALL_OUT_List = apply(X=U, MARGIN=1, FUN=function(x1) f(x = x1, t="MOC"))
# Total communication count between all B-parties
TOTAL_List = SMS_IN_List+SMS_OUT_List+CALL_IN_List+CALL_OUT_List
# Create pivot dataframe: B-Party and communication count details
df_pivot <- data.frame(U,CALL_IN_List,CALL_OUT_List,SMS_IN_List,SMS_OUT_List,TOTAL_List)
colnames(df_pivot) <- c("B_Party", "IN", "OUT", "SMS_IN", "SMS_OUT", "Total") # give column names
df_pivot <- df_pivot[order(-df_pivot$Total),] # Note the (-) sign: sort in descending order of total communication
# Output pivot details to a file
outfile = readline(prompt="Output file: ")
sheet = readline(prompt="Sheet name: ")
hs <- createStyle(textDecoration = "BOLD", fontColour = "#FFFFFF", fontSize=12, fontName="Arial Narrow", fgFill = "#4F80BD")
write.xlsx(x=df_pivot, file=outfile, sheetName = sheet, headerStyle=hs) |
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/strataMethods.R
\docType{methods}
\name{strata}
\alias{strata}
\alias{strata,genind-method}
\alias{strata,genlight-method}
\alias{strata<-}
\alias{strata<-,genind-method}
\alias{strata<-,genlight-method}
\alias{nameStrata}
\alias{nameStrata,genind-method}
\alias{nameStrata,genlight-method}
\alias{nameStrata<-}
\alias{nameStrata<-,genind-method}
\alias{nameStrata<-,genlight-method}
\alias{splitStrata}
\alias{splitStrata,genind-method}
\alias{splitStrata,genlight-method}
\alias{splitStrata<-}
\alias{splitStrata<-,genind-method}
\alias{splitStrata<-,genlight-method}
\alias{addStrata}
\alias{addStrata,genind-method}
\alias{addStrata,genlight-method}
\alias{addStrata<-}
\alias{addStrata<-,genind-method}
\alias{addStrata<-,genlight-method}
\title{Access and manipulate the population strata for genind or genlight objects.}
\usage{
strata(x, formula = NULL, combine = TRUE, value)
strata(x) <- value
nameStrata(x, value)
nameStrata(x) <- value
splitStrata(x, value, sep = "_")
splitStrata(x, sep = "_") <- value
addStrata(x, value, name = "NEW")
addStrata(x, name = "NEW") <- value
}
\arguments{
\item{x}{a genind or genlight object}
\item{formula}{a nested formula indicating the order of the population
strata.}
\item{combine}{if \code{TRUE} (default), the levels will be combined according to the
formula argument. If it is \code{FALSE}, the levels will not be combined.}
\item{value}{a data frame OR vector OR formula (see details).}
\item{sep}{a \code{character} indicating the character used to separate
hierarchical levels. This defaults to "_".}
\item{name}{an optional name argument for use with addStrata if supplying
a vector. Defaults to "NEW".}
}
\description{
The following methods allow the user to quickly change the strata of a genind
or genlight object.
}
\details{
\subsection{Function Specifics}{ \itemize{ \item \strong{strata()} -
Use this function to view or define population stratification of a
\linkS4class{genind} or \linkS4class{genlight} object. \item
\strong{nameStrata()} - View or rename the different levels of strata.
\item \strong{splitStrata()} - Split strata that are combined with a common
separator. This function should only be used once during a workflow.
\itemize{ \item \emph{Rationale:} It is often difficult to import files
with several levels of strata as most data formats do not allow unlimited
population levels. This is circumvented by collapsing all population strata
into a single population factor with a common separator for each
observation. } \item \strong{addStrata()} - Add levels to your population
strata. This is ideal for adding groups defined by
\code{\link{find.clusters}}. You can input a data frame or a vector, but if
you put in a vector, you have the option to name it. }}
\subsection{Argument Specifics}{
These functions allow the user to seamlessly carry all possible population
stratification with their \linkS4class{genind} or \linkS4class{genlight}
object. Note that there are two ways of performing all methods: \itemize{
\item modifying: \code{strata(myData) <- myStrata} \item preserving:
\code{myNewData <- strata(myData, value = myStrata)} } They essentially do
the same thing except that the modifying assignment method (the one with
the "\code{<-}") will modify the object in place whereas the non-assignment
method will preserve the original object (unless you overwrite it). Due to
convention, everything right of the assignment is termed \code{value}. To
avoid confusion, here is a guide to the argument \strong{\code{value}} for
each function: \itemize{ \item \strong{strata()} \code{value = }a
\code{\link{data.frame}} that defines the strata for each individual in the
rows. \item \strong{nameStrata()} \code{value = }a \code{\link{vector}} or
a \code{\link{formula}} that will define the names. \item
\strong{splitStrata()} \code{value = }a \code{\link{formula}} argument with
the same number of levels as the strata you wish to split. \item
\strong{addStrata()} \code{value = }a \code{\link{vector}} or
\code{\link{data.frame}} with the same length as the number of individuals
in your data. }}
\subsection{Details on Formulas}{
The preferred use of these functions is with a \code{\link{formula}}
object. Specifically, a hierarchical formula argument is used to assign the
levels of the strata. An example of a hierarchical formula would
be:\tabular{r}{ \code{~Country/City/Neighborhood}} This convention was
chosen as it becomes easier to type and makes intuitive sense when defining
a \code{\link{hierarchy}}. Note: it is important to use hiearchical
formulas when specifying hierarchies as other types of formulas (eg.
\code{~Country*City*Neighborhood}) will give incorrect results.}
}
\examples{
# let's look at the microbov data set:
data(microbov)
microbov
# We see that we have three vectors of different names in the 'other' slot.
# ?microbov
# These are Country, Breed, and Species
names(other(microbov))
# Let's set the strata
strata(microbov) <- data.frame(other(microbov))
microbov
# And change the names so we know what they are
nameStrata(microbov) <- ~Country/Breed/Species
\dontrun{
# let's see what the strata looks like by Species and Breed:
head(strata(microbov, ~Breed/Species))
# If we didn't want the last column combined with the first, we can set
# combine = FALSE
head(strata(microbov, ~Breed/Species, combine = FALSE))
#### USING splitStrata ####
# For the sake of example, we'll imagine that we have imported our data set
# with all of the stratifications combined.
setPop(microbov) <- ~Country/Breed/Species
strata(microbov) <- NULL
# This is what our data would look like after import.
microbov
# To set our strata here, we need to use the functions strata and splitStrata
strata(microbov) <- data.frame(x = pop(microbov))
microbov # shows us that we have "one" level of stratification
head(strata(microbov)) # all strata are separated by "_"
splitStrata(microbov) <- ~Country/Breed/Species
microbov # Now we have all of our strata named and split
head(strata(microbov)) # all strata are appropriately named and split.
}
}
\seealso{
\code{\link{setPop}} \code{\link{genind}}
\code{\link{as.genind}}
}
\author{
Zhian N. Kamvar
}
| /man/strata-methods.Rd | no_license | thibautjombart/adegenet | R | false | true | 6,376 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/strataMethods.R
\docType{methods}
\name{strata}
\alias{strata}
\alias{strata,genind-method}
\alias{strata,genlight-method}
\alias{strata<-}
\alias{strata<-,genind-method}
\alias{strata<-,genlight-method}
\alias{nameStrata}
\alias{nameStrata,genind-method}
\alias{nameStrata,genlight-method}
\alias{nameStrata<-}
\alias{nameStrata<-,genind-method}
\alias{nameStrata<-,genlight-method}
\alias{splitStrata}
\alias{splitStrata,genind-method}
\alias{splitStrata,genlight-method}
\alias{splitStrata<-}
\alias{splitStrata<-,genind-method}
\alias{splitStrata<-,genlight-method}
\alias{addStrata}
\alias{addStrata,genind-method}
\alias{addStrata,genlight-method}
\alias{addStrata<-}
\alias{addStrata<-,genind-method}
\alias{addStrata<-,genlight-method}
\title{Access and manipulate the population strata for genind or genlight objects.}
\usage{
strata(x, formula = NULL, combine = TRUE, value)
strata(x) <- value
nameStrata(x, value)
nameStrata(x) <- value
splitStrata(x, value, sep = "_")
splitStrata(x, sep = "_") <- value
addStrata(x, value, name = "NEW")
addStrata(x, name = "NEW") <- value
}
\arguments{
\item{x}{a genind or genlight object}
\item{formula}{a nested formula indicating the order of the population
strata.}
\item{combine}{if \code{TRUE} (default), the levels will be combined according to the
formula argument. If it is \code{FALSE}, the levels will not be combined.}
\item{value}{a data frame OR vector OR formula (see details).}
\item{sep}{a \code{character} indicating the character used to separate
hierarchical levels. This defaults to "_".}
\item{name}{an optional name argument for use with addStrata if supplying
a vector. Defaults to "NEW".}
}
\description{
The following methods allow the user to quickly change the strata of a genind
or genlight object.
}
\details{
\subsection{Function Specifics}{ \itemize{ \item \strong{strata()} -
Use this function to view or define population stratification of a
\linkS4class{genind} or \linkS4class{genlight} object. \item
\strong{nameStrata()} - View or rename the different levels of strata.
\item \strong{splitStrata()} - Split strata that are combined with a common
separator. This function should only be used once during a workflow.
\itemize{ \item \emph{Rationale:} It is often difficult to import files
with several levels of strata as most data formats do not allow unlimited
population levels. This is circumvented by collapsing all population strata
into a single population factor with a common separator for each
observation. } \item \strong{addStrata()} - Add levels to your population
strata. This is ideal for adding groups defined by
\code{\link{find.clusters}}. You can input a data frame or a vector, but if
you put in a vector, you have the option to name it. }}
\subsection{Argument Specifics}{
These functions allow the user to seamlessly carry all possible population
stratification with their \linkS4class{genind} or \linkS4class{genlight}
object. Note that there are two ways of performing all methods: \itemize{
\item modifying: \code{strata(myData) <- myStrata} \item preserving:
\code{myNewData <- strata(myData, value = myStrata)} } They essentially do
the same thing except that the modifying assignment method (the one with
the "\code{<-}") will modify the object in place whereas the non-assignment
method will preserve the original object (unless you overwrite it). Due to
convention, everything right of the assignment is termed \code{value}. To
avoid confusion, here is a guide to the argument \strong{\code{value}} for
each function: \itemize{ \item \strong{strata()} \code{value = }a
\code{\link{data.frame}} that defines the strata for each individual in the
rows. \item \strong{nameStrata()} \code{value = }a \code{\link{vector}} or
a \code{\link{formula}} that will define the names. \item
\strong{splitStrata()} \code{value = }a \code{\link{formula}} argument with
the same number of levels as the strata you wish to split. \item
\strong{addStrata()} \code{value = }a \code{\link{vector}} or
\code{\link{data.frame}} with the same length as the number of individuals
in your data. }}
\subsection{Details on Formulas}{
The preferred use of these functions is with a \code{\link{formula}}
object. Specifically, a hierarchical formula argument is used to assign the
levels of the strata. An example of a hierarchical formula would
be:\tabular{r}{ \code{~Country/City/Neighborhood}} This convention was
chosen as it becomes easier to type and makes intuitive sense when defining
a \code{\link{hierarchy}}. Note: it is important to use hiearchical
formulas when specifying hierarchies as other types of formulas (eg.
\code{~Country*City*Neighborhood}) will give incorrect results.}
}
\examples{
# let's look at the microbov data set:
data(microbov)
microbov
# We see that we have three vectors of different names in the 'other' slot.
# ?microbov
# These are Country, Breed, and Species
names(other(microbov))
# Let's set the strata
strata(microbov) <- data.frame(other(microbov))
microbov
# And change the names so we know what they are
nameStrata(microbov) <- ~Country/Breed/Species
\dontrun{
# let's see what the strata looks like by Species and Breed:
head(strata(microbov, ~Breed/Species))
# If we didn't want the last column combined with the first, we can set
# combine = FALSE
head(strata(microbov, ~Breed/Species, combine = FALSE))
#### USING splitStrata ####
# For the sake of example, we'll imagine that we have imported our data set
# with all of the stratifications combined.
setPop(microbov) <- ~Country/Breed/Species
strata(microbov) <- NULL
# This is what our data would look like after import.
microbov
# To set our strata here, we need to use the functions strata and splitStrata
strata(microbov) <- data.frame(x = pop(microbov))
microbov # shows us that we have "one" level of stratification
head(strata(microbov)) # all strata are separated by "_"
splitStrata(microbov) <- ~Country/Breed/Species
microbov # Now we have all of our strata named and split
head(strata(microbov)) # all strata are appropriately named and split.
}
}
\seealso{
\code{\link{setPop}} \code{\link{genind}}
\code{\link{as.genind}}
}
\author{
Zhian N. Kamvar
}
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/jackse.R
\name{jackse}
\alias{jackse}
\title{Jackknife standard errors}
\usage{
jackse(x, theta, bycol = TRUE, ...)
}
\arguments{
\item{x}{A vector or matrix to be summarized.}
\item{theta}{The summary statistic to be used. For example, \code{theta = mean} calculates the
jackknife standard errors for the mean.}
\item{bycol}{Logical indicating if the data are in rows or columns when x is a matrix. The
default is TRUE.}
\item{...}{Other parameters to be passed to the summary statistic function specified by
\code{theta}.}
}
\value{
Jackknife standard errors.
}
\description{
Calculate jackknife standard errors.
}
\author{
Ander Wilson
}
| /man/jackse.Rd | no_license | my1120/smurf | R | false | true | 725 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/jackse.R
\name{jackse}
\alias{jackse}
\title{Jackknife standard errors}
\usage{
jackse(x, theta, bycol = TRUE, ...)
}
\arguments{
\item{x}{A vector or matrix to be summarized.}
\item{theta}{The summary statistic to be used. For example, \code{theta = mean} calculates the
jackknife standard errors for the mean.}
\item{bycol}{Logical indicating if the data are in rows or columns when x is a matrix. The
default is TRUE.}
\item{...}{Other parameters to be passed to the summary statistic function specified by
\code{theta}.}
}
\value{
Jackknife standard errors.
}
\description{
Calculate jackknife standard errors.
}
\author{
Ander Wilson
}
|
#install.packages('rsconnect')
#library(rsconnect)
#install.packages("shinyjs")
options(scipen=999)
# imports -----------------------------------------------------------------
library(shiny)
library(shinyjs)
library(leaflet)
library(RColorBrewer)
library(scales)
library(lattice)
library(dplyr)
library(ggplot2)
library(plotly)
library(data.table)
library(lubridate)
library(leaflet.extras)
library(magrittr) # chain operators, e.g. to "pipe" a value forward
#library(plyr)
library(tidyverse)
library(DT)
library(knitr)
library(maps)
library(rgdal)
library(ggmap)
library(tmap)
library(sp)
library(tmap)
library(sf)
library(stars)
library(spData)
library(classInt)
library(lattice)
library(grid)
library(pals)
# prajwal's data fetch ----------------------------------------------------------
# Download the data from https://data.cityofnewyork.us/api/views/3q43-55fe/rows.csv?accessType=DOWNLOAD
# Alternate link: https://data.cityofnewyork.us/Social-Services/Rat-Sightings/3q43-55fe, click Export -> CSV
rat_sightings <- read.csv("https://data.cityofnewyork.us/api/views/3q43-55fe/rows.csv?accessType=DOWNLOAD")
#rat_sightings <- read.csv("data/Rat_Sightings.csv")
rat_sightings <- rat_sightings %>% filter(!(Status=='Open'))
rat_sightings <- rat_sightings %>% filter(!(Status=='Draft'))
rat_sightings <- rat_sightings %>% filter(!(Borough=='Unspecified'))
rat_sightings$latitude <- rat_sightings$Latitude
rat_sightings$longitude <- rat_sightings$Longitude
#set.seed(100)
#c("BROOKLYN", "QUEENS","STATEN ISLAND")
#rat_sightings_buroughs <- c("BROOKLYN", "QUEENS","STATEN ISLAND")
# PRATISHTA'S DATA FETCH -------------------------------------------------------
# read in the main csv file
rat_data<-read.csv("data/rat_data.csv")
rat_data <- rat_data %>%
mutate(Borough = str_to_title(rat_data$Borough))
tonnage_data<-read.csv("data/dsny_boro_tonnage.csv", stringsAsFactors = FALSE)
ton_date <- tonnage_data %>%
mutate(MONTH = paste(MONTH, " / 01")) %>%
mutate(MONTH = as.Date(MONTH, format = '%Y / %m / %d')) %>%
filter(MONTH > as.Date('2020-01-01', '%Y-%m-%d'), MONTH < as.Date('2021-03-01', '%Y-%m-%d')) %>%
arrange(desc(MONTH))
rat_date <- rat_data %>%
mutate(Created.Date = as.Date(Created.Date, "%m/%d/%Y")) %>%
mutate(Created.Date = as.character(Created.Date)) %>%
mutate(Created.Date = substr(Created.Date, 1, 8)) %>%
mutate(Created.Date = paste(Created.Date, '01')) %>%
mutate(Created.Date = as.Date(Created.Date, "%Y-%m-%d")) %>%
group_by(Created.Date, Borough) %>%
tally() %>%
filter(Created.Date > as.Date('2020-01-01', '%Y-%m-%d'), Created.Date < as.Date('2021-03-01', '%Y-%m-%d')) %>%
arrange(desc(Created.Date))
rat_ton_date <- merge(rat_date, ton_date, by.x = c("Created.Date", "Borough"), by.y = c("MONTH", "BOROUGH")) %>%
mutate(rate = n / (REFUSETONSCOLLECTED / 100))
# END OF PRATISHTA'S DATA FETCH ------------------------------------------------
# PRATISHTA'S CODE -------------------------------------------------------------
# community district conversion functions
convertBoroCDToDistrict <- function(borocd) {
sapply(borocd, function(borocd) {
boro_ch = as.character(borocd)
boro_n = substr(boro_ch, 1, 1)
cd_n = substr(boro_ch, 2, 3)
boro = case_when (boro_n == '1' ~ 'MN',
boro_n == '2' ~ 'BX',
boro_n == '3' ~ 'BK',
boro_n == '4' ~ 'QW',
boro_n == '5' ~ 'SI'
)
ans <- paste(boro, cd_n, sep="")
return (ans)
})
}
convertToShpDistrict <- function(com_district) {
sapply(com_district, function(com_district) {
split = strsplit(com_district, " ")
boro = case_when (str_to_lower(split[[1]][2]) == 'brooklyn' ~ 'BK',
str_to_lower(split[[1]][2]) == 'manhattan' ~ 'MN',
str_to_lower(split[[1]][2]) == 'queens' ~ 'QW',
str_to_lower(split[[1]][2]) == 'staten' ~ 'SI',
str_to_lower(split[[1]][2]) == 'bronx' ~ 'BX'
);
ans <- paste(boro, split[[1]][1], sep="")
return (ans)
})
}
# reading in data and modify string format of community district column
full_tonnage <-read.csv("sanitation_data/dsny_full_tonnage.csv", stringsAsFactors = FALSE)
full_tonnage <- full_tonnage %>%
mutate(district = paste(full_tonnage$COMMUNITYDISTRICT, str_to_upper(full_tonnage$BOROUGH)))
district = paste(full_tonnage$COMMUNITYDISTRICT, str_to_upper(full_tonnage$BOROUGH))
# creating data to be mapped
ton_map <- full_tonnage %>%
mutate(community_district = convertToShpDistrict(district)) %>%
group_by(community_district) %>%
summarise(total_ton = sum(REFUSETONSCOLLECTED))
ton_map
community_district <- paste(rat_data$Community.Board, str_to_upper(rat_data$Community.Board))
rat_map <- rat_data %>%
mutate(community_district = convertToShpDistrict(community_district)) %>%
group_by(community_district) %>%
tally()
rat_map
rat_borough <- rat_data %>%
group_by(Borough) %>%
tally()
rat_borough
ton_boro <- tonnage_data %>%
group_by(BOROUGH) %>%
summarise(total_ton = sum(REFUSETONSCOLLECTED))
ton_boro
rat_ton <- left_join(rat_borough, ton_boro, by = c("Borough" = "BOROUGH"))
rat <- ggplot(rat_ton, aes(y=n, x=Borough, fill = Borough)) +
geom_bar(position="dodge", stat="identity")
ton <- ggplot(rat_ton, aes(y=total_ton, x=Borough, fill = Borough)) +
geom_bar(position="dodge", stat="identity")
# map data
nyc <- readOGR("sanitation_data/CommunityDistricts/.", "geo_export_d81daad1-2b49-44c3-81d4-72436a58def3")
nyc_sp <- spTransform(nyc, CRS("+proj=longlat +datum=WGS84"))
nyc_sp@data <- nyc_sp@data %>%
mutate(community_district = convertBoroCDToDistrict(boro_cd))
nyc_sp@data
nyc_sp@data <- left_join(nyc_sp@data, rat_map)
nyc_sp@data
nyc_sp@data <- left_join(nyc_sp@data, ton_map)
nyc_sp@data
# END OF PRATISHTA'S CODE-------------------------------------------------------
# shiny code
# brendan's imports and code -------------------------------------------------------
library(ggplot2)
library(ggthemes)
library(gridExtra)
library(dplyr)
library(readr)
library(leaflet)
library(leaflet.extras)
library(magrittr)
library(dplyr)
library(tidyr)
library(wordcloud)
library(png)
library(ggwordcloud)
library(tidytext)
library(readr)
library(png)
#setwd("~/Bayesian")
open <- read.csv("data/Open_Restaurant_Applications.csv")
inspection <- read.csv("data/DOHMH_New_York_City_Restaurant_Inspection_Results.csv")
rat_311 <- read.csv("data/311_Service_Requests_from_2010_to_Present.csv")
restaurant_borough <- count(open, Borough)
names(restaurant_borough)[names(restaurant_borough) == "n"] <- "count"
rat_borough <- count(rat_311, Borough)
borough <- c("Bronx", "Brooklyn", "Manhattan", "Queens", "Staten Island", "none")
rat_borough <- cbind(rat_borough, borough) %>% filter(borough!= "none")
rat_borough <- select(rat_borough, borough, n)
names(rat_borough)[names(rat_borough) == "n"] <- "count"
names(rat_borough)[names(rat_borough) == "borough"] <- "Borough"
inspection_bc <- inspection %>% filter(GRADE == "B" | GRADE == "C")
inspection_count_2020 <- count(inspection_bc, BORO)
names(inspection_count_2020)[names(inspection_count_2020) == "n"] <- "count"
names(inspection_count_2020)[names(inspection_count_2020) == "BORO"] <- "Borough"
street_seating <- filter(open, Approved.for.Roadway.Seating == "yes")
count_street <- count(street_seating, Borough)
names(count_street)[names(count_street) == "n"] <- "count"
sidewalk_seating <- filter(open, Approved.for.Sidewalk.Seating == "yes")
count_sidewalk <- count(sidewalk_seating, Borough)
names(count_sidewalk)[names(count_sidewalk) == "n"] <- "count"
manhattan_311 <- filter(rat_311, Borough == "MANHATTAN")
#manhattan_311 <- read.csv("manhattan311.csv")
manhattan_open <- read.csv("data/manhattan open restaurants.csv")
manhattan_311 <- filter(manhattan_311, manhattan_311$Complaint.Type=="Rodent")
manhattan_311 <- data.frame(manhattan_311$Latitude, manhattan_311$Longitude, manhattan_311$Incident.Address, manhattan_311$Created.Date, manhattan_311$Descriptor)
icon <- makeIcon(iconUrl= "https://cdn3.iconfinder.com/data/icons/farm-animals/128/mouse-512.png", iconWidth=25, iconHeight = 20)
#manhattan_311b <- read.csv("manhattan311.csv")
#manhattan_311b <- read.csv("https://nycopendata.socrata.com/api/views/erm2-nwe9/rows.csv?accessType=DOWNLOAD")
# code for making wordcloud (not used in live version for processing power reasons) -----------------------------------------------
# manhattan_311b <- filter(rat_311, Borough == "MANHATTAN")
# nrow(manhattan_311b)
#
# manhattan_311b <- manhattan_311b %>% filter(Complaint.Type != "Noise = Street/Sidewalk" & Complaint.Type != "Noise - Residential" & Complaint.Type != "HEAT/HOT WATER" & Complaint.Type != "Illegal Parking" & Complaint.Type != "Non-Emergency Police Matter" & Complaint.Type != "Noise" & Complaint.Type != "Noise - Vehicle" & Complaint.Type != " Noise - Commercial")
# descriptors <- manhattan_311b %>% select(Descriptor)
# descriptors_fix <- as.character(descriptors$Descriptor)
# text_df <- tibble(line = 1:length(descriptors_fix), text = descriptors_fix)
# descriptors <- text_df %>% unnest_tokens(word, text)
#
# descriptors <- count(descriptors, word)
# descriptors2 <- filter(descriptors, n > 2000)
# col <- c(ifelse(descriptors2$word == "pests" | descriptors2$word == "rat" | descriptors2$word == "sighting" | descriptors2$word == "rodents", "red", "black"))
# descriptors3 <- cbind(descriptors2, col)
# descriptors3 <- filter(descriptors3, word != "n" & word != "a" & word != "not" & word != "business" & word != "no" & word!= "compliance" & word != "or" & word != "in" & word != "of" & word!= "to" & word!= "non" & word!= "on" & word != "has" & word!= "for")
# #setwd("~/Bayesian")
# img <- readPNG("data/rat.png")
# #img <- icon
# descriptors3 <- descriptors3 %>% filter(word != "loud" & word!= "music" & word != "party")
# set.seed(14)
# (wordcloud1 <- ggplot(descriptors3, aes(label=word, size=n, color=col)) + geom_text_wordcloud_area(mask=img, rm_outside = TRUE) + scale_size_area(max_size=5) + theme_classic())
# (wordcloud2 <- ggplot(descriptors3, aes(label=word, size=n, color=col)) + geom_text_wordcloud_area() + scale_size_area())
# user interface for setting layout of plots ----------------------------------------------------------
# sliders and interactive map ---------------------------------------------
rat_sightings_buroughs <- as.character(unique(unlist(rat_sightings$Borough)))
rat_sightings_case_status <- as.character(unique(unlist(rat_sightings$Status)))
ui <- fluidPage(
# header description ------------------------------------------------------
tags$head(
# Note the wrapping of the string in HTML()
tags$style(HTML("
.row {
margin-left: 0;
margin-right:0;
}"))
),
fluidRow(align = "center",
h1("Rats and NYC: Exploratory Visualization"),
strong("Data Visualization (QMSS - G5063) Final Project, Spring 2021"),
br(),
em("Group N: Brendan Mapes, Prajwal Seth, and Pratishta Yerakala"),
h3(a("Link to code and process book", href="https://github.com/QMSS-G5063-2021/Group_N_NYCdata")),
br(),br(),br(),
p("In this project, we will explore in detail New York City's rat problem.
New York has always dealt with a relatively high population of rats, as do many
other large metropolitan areas. However, since the beginning of the COVID-19 pandemic
rat sightings have been on the rise. Rats are being seen more often, during new times of day,
and are acting more aggressive. Through the following visualizations we hope to find some explanation
for this recent uptick in rat sightings. The way restaurants and residents handle their trash plays
a large role in the survival and behavior of rats in the city. So through exploration of city
sanitation data, restaurant registration data, and 311 calls in the city, we hope to find some
potential explanations as to why New York's rat problem has gotten so bad."),),
br(),br(),br(),
# prajwal's description ---------------------------------------------------
# fluidRow(
# align = "center",
# headerPanel("Hello 1!"),
# p("p creates a paragraph of text."),
# p("A new p() command starts a new paragraph. Supply a style attribute to change the format of the entire paragraph.", style = "font-family: 'times'; font-si16pt"),
# strong("strong() makes bold text."),
# em("em() creates italicized (i.e, emphasized) text."),
# br(),
# code("code displays your text similar to computer code"),
# div("div creates segments of text with a similar style. This division of text is all blue because I passed the argument 'style = color:blue' to div", style = "color:blue"),
# br(),
# p("span does the same thing as div, but it works with",
# span("groups of words", style = "color:blue"),
# "that appear inside a paragraph."),
# ),
fluidRow(
align = "center",
style='margin-left:0px; margin-right: 0px;',
h2("Interactive map of rodent complaints in NYC since 2010"),
h3("Prajwal Seth"),
br(),
),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings (automatically updated daily)", href="https://data.cityofnewyork.us/Social-Services/Rat-Sightings/3q43-55fe"),
br(),
h4("Background:"),
p("In this section, I have visualized rodent complaints from 2010 till today submitted to the NYC311 portal.
Feel free to play around with the provided filters for number of samples,
years, boroughs, and case status (however, due to processing constraints on shinyapps.io,
the website will crash if you set the number of samples too high). The map on the left will dynamically update as you
change the filters (or will not update if there is no data to display after the filters are applied).
The plot for the trend in complaint status also updates according to the
rat complaints that are visible in the map. Upon zooming into the map, you will see that the color of the
marker for each complaint is set according to the complaint status (refer to the legend of the map).
Also provided is a tooltip displaying the complaint's created date, closed date, address, status, and
location type. There is a layer of heat added to the map, with the intensity of the heat
being calculated based on the number of rat sightings in the area."),
),
),
# sliders and map etc -----------------------------------------------------
fluidRow(
sidebarLayout(position = "right",
sidebarPanel(width = 6,
sliderInput("num_sample", label = h4("Select number of random samples"), min = 1,
max = nrow(rat_sightings), value = 1000, step = 1000),
sliderInput("year_input", label = h4("Select years"), min = 2010,
max = 2021, value = c(2010, 2021), step = 1, format = "####"),
#selected = rat_sightings_buroughs[1:length(multiInput)])
#selected = rat_sightings_buroughs,
selectizeInput("burough_input", label=h4("Select boroughs"), choices =rat_sightings_buroughs, multiple = TRUE, selected = rat_sightings_buroughs),
selectizeInput("case_status", label=h4("Select status"), choices =rat_sightings_case_status, multiple = TRUE, selected = rat_sightings_case_status),
#plotlyOutput("cityViz", height = 300),
plotlyOutput("yearViz", height = 250),
#plotlyOutput("locationViz", height = 220),
#plotlyOutput("locationViz", height = 300),
),
mainPanel(width = 6, style='margin-top:40px;',
leafletOutput("map", height = 825),
),
),
# PRATISHTA'S WRITEUP --------------------------------------------------------
fluidRow(
align = "center",
style='margin-left:0px; margin-right: 0px;',
h2("Rat Sightings and Sanitation Waste by Borough"),
h3("Pratishta Yerakala"),
br(),
),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings", href="https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9"), h6("filtered for rodent sightings between Feb 2020 and Feb 2021.")),
h5(a("DSNY Monthly Tonnage", href="https://data.cityofnewyork.us/City-Government/DSNY-Monthly-Tonnage-Data/ebb7-mvp5"), h6("filtered for months between Feb 2020 and Feb 2021.")),
h5(a("NYC Community Districts Shapefile", href="https://data.cityofnewyork.us/City-Government/Community-Districts/yfnk-k7r4")),
),
div(class='padding',
h4("Background:"),
p("The large rodent population in New York City is no secret. Rats have been
associated with the city for a long time whether it's from the famous", a('pizza rat', href='https://knowyourmeme.com/memes/pizza-rat'),
"or to the rising concern from residents who have noticed changes since the COVID-19 pandemic.
Many businesses and normally occurring procedures have been slowed down or
halted completely. One such example in particular with the Department of
Sanitation of NY (DSNY) where limited resources and budget cuts since the
pandemic have caused an", a("increased amount of litter and waste production", href="https://patch.com/new-york/new-york-city/city-state-leaders-decry-sanitation-setback-trash-piles"),
"."),
)
),
fluidRow(
align = "center",
div(class='padding',
h4(align = "left", "Visualizations:"),
),
# descriptive charts
h3("Total Number of Rat Sightings (2020-2021)"),
h6("Chart 1"),
plotlyOutput("pratishta5", width = "50%"),
br(),
h3("Total Waste Produced (2020-2021)"),
h6("Chart 2"),
plotlyOutput("pratishta4", width = "50%"),
br(),
p(class = "padding", align = "left", "We can see in Chart 1 that Brooklyn produces the most tons of waste followed by
Queens, and then by Bronx and Manhattan. Staten Island is last with the
least amount of waste. Chart 2 shows the number of rat sightings per
borough and naturally, we have Brooklyn at the top with around 6,000
sightings. But Instead of Queens, Manhattan follows with most rat s
ightings. Then Queens and Bronx. From here it seems that Staten Island and
Bronx are boroughs that have some what proportional sightings to waste
produced. However, though Queens produces a lot of waste, it does not have
nearly the same rate of rat sightings. Conversely, Manhattan doesn't quite
produce the same amount of waste as Queens but seems to have far more rat
sightings. Brooklyin is consistenly infested."),
# time series charts
h3("Waste per Month (2020-2021)"),
h6("Chart 3"),
plotlyOutput("pratishta1", width = "70%"),
br(),
h3("Rat Sightings per Month (2020-2021)"),
h6("Chart 4"),
plotlyOutput("pratishta2", width = "70%"),
br(),
h3("Rate of Rats per Waste Ton per Month"),
h6("Chart 5"),
plotlyOutput("pratishta3", width = "70%"),
br(),
p(class = "padding", align = "left", "Charts 3 to 5 show a time series line graphs. Chart 3 shows the tons of
waste per month generated by each borough. It seems that around March and
April of 2020 there was a change in the trend. Though the waste was
rising, it flattened out - or even fell like with Manhattan for example
- between March and April of 2020. But after april there was a more mellow
rise and then gentle decline closer to Fall of 2020 and early 2021. This
exploratory chart is good to possibly check out Manhattan and why the
waste production went down. Perhaps for restaurants closing?"),
p(class = "padding", align = "left", "Chart 4 shows that all boroughs saw an increase in rat sightings,
especially Brooklyn. It did seem to peak around summer of 2020 and decline
again to almost normal rates. These sightings might be due to the
sanitation departments' limits as mentioned earlier."),
p(class = "padding", align = "left", "Chart 5 looks at the 'rate' at which a rat sighting is possible for every
kilo-ton of waste produced in each borough per month. It seems that these
rates also follow a similar path of an increase around April 2020 and then
a peak in summer of 2020 and then a mellow decline. However Manhattan's
rate shot up for early 2021 with around 0.75 (sightings per ton of waste)
to 1.5 (sightings per ton of waste). Perhaps this could be due to frequent
testings, vaccinations, and re-opening of restaurants (producing more
waste)?"),
# maps
h3("Number of Rat Sightings per Month"),
h6("Chart 6"),
),
fluidRow(
tmapOutput("pratishta7", width = "100%"),
br(),
),
fluidRow(align = "center",
h3("Waste Produced in Tons"),
h6("Chart 7"),
),
fluidRow(
tmapOutput("pratishta8", width = "100%"),
br(),
),
fluidRow(
align = "center",
h3("Rat Sightings and Waste Produced By Community District"),
h6("Chart 8"),
plotOutput("pratishta9", width = "60%"),
br(),
p(class = "padding", align = "left", "Here we show choropleth maps of NYC by
community district. Community district as the geographical feature here is
because DSNY also records their information with that feature. It seems
that there is one community districte that's seeing a severe rise in rat
sightings (BK03). It may be worth it to take a closer look at that
particular district if analysis or further studie are done."),
p(class = "padding", align = "left", "As expected it looks like Queens
produces a massive amount of waste and the particular districts are
highlighted in deep red (100,000 - 120,000 tons category). And though
Brooklyn also produces a lot of waste it seems to be spread out amongst
the community districts."),
p(class = "padding", align = "left", "Chart 8 depicts a bivariate choropleth
map of the community districts. This make-shift rendering of a map using
the tm_map package is from", a("this currently open issue from the tmap GitHub
repsitory", href="https://github.com/mtennekes/tmap/issues/183#issuecomment-670554921"),
". As the districts become more blue in color, the more right sightings
that have been reported in that district. The more purple a district, the
more amount of waste produced there. As the color goes to indigo, there's
a high rat sighting and a high waste production (a high high category if
further spatial dependence analysis was conducted). Light grey represents
few rat sightings and little waste production (a 'low low' again for further
spatial research). This map demonstrates that hough there are high rat
sightings in downtown Manhattan and parts of Brooklyn, it's not
necessarily tied to waste production. In the same way, the outer boroughs
have a huge waste production but not many rat sightings. But there are
some districts (indigo) that do exhibit both features in high amounts.
This exploratory data visulaization provides the insight to further look
in those districts.")
),
# Brendan's writeup -------------------------------------------------
fluidRow(
align = "center",
h2("Rat Sightings and Restaurants by Borough"),
h3("Brendan Mapes"),
br(),
),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings", href="https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9"), h6("Filtered to 311 calls from Jan 1 2020 to Dec 31 2020, of complaint type βrodentβ.")),
h5(a("Open Restaurants", href="https://data.cityofnewyork.us/Transportation/Open-Restaurant-Applications/pitm-atqc/data"),),
),
div(class='padding',
h4("Background:"),
p("It is well documented that the COVID-19 pandemic has led to a rise in rat sightings throughout New York
City. It is also well known that the pandemic has been especially hard on the restaurant industry. Hit
especially hard by stay-at-home orders and social distancing mandates, restaurants have been forced to
innovate their operations. Now more than ever, restaurants are serving customers outside in the
streets. We suspect this change in the way restaurants do their business may be contributing to the
increase in rat sightings. We explore this possibility a bit further in the next few visualizations."),
),
div(class='padding',
h4("Visualizations:"),
),
),
# descriptive charts
# fluidRow(align = "center",
# br(),
# h3("NYC311 rodent complaints in 2020 and number of restaurants by type"),
# h6("Chart 9"),br(),
# plotOutput("brendan_chart1", width = "80%"),
# plotOutput("brendan_chart2", width = "80%"),
# br(),
# br(),
# p(class = "padding", align = "left", "In all four figures, we can see that Manhattan is far above the rest of the boroughs in restaurants
# approved for outdoor dining, in sidewalk and street dining. However, it is Brooklyn that is far above the
# rest of the boroughs in rodent reports in 2020. This suggests that perhaps another factor is contributing
# to the rat problem in the Brooklyn borough. If restaurants were fully to blame for itβs rat problem, we
# would expect to see it having high numbers of restaurants approved for outdoor street and sidewalk
# dining, a number comparable to the borough of Manhattan.
#
# The first bar plot displays the number of rodent related 311 reports in the year 2020 by borough.
# Brooklyn leads the way with well over 10,000 rodent related calls in the year, while the next closest
# borough, Manhattan, only has about 8,000 rodent related calls in the year. In the bar plots related to
# restaurants, we see Manhattan leads the way across the board. In the restaurants with outdoor dining,
# sidewalk and street dining, Manhattan has twice as many restaurants than any other borough. Because
# of this vast difference in the number of restaurants in Manhattan compared to the other boroughs, we
# will narrow our focus to Manhattan in the next visualization."),),
fluidRow(align = "center",
br(),
br(),
br(),
# code for generating these plots
# title: "brendan2"
# author: "Brendan Mapes"
# date: "4/17/2021"
# output: html_document
# ---
#
# ```{r setup, include=TRUE}
# knitr::opts_chunk$set(echo = TRUE)
# library(ggplot2)
# library(ggthemes)
# library(gridExtra)
# library(dplyr)
# library(plotly)
# open <- read.csv("Open_Restaurant_Applications.csv")
# inspection <- read.csv("DOHMH_New_York_City_Restaurant_Inspection_Results.csv")
# rat_311 <- read.csv("311_Service_Requests_from_2010_to_Present.csv")
# restaurant_borough <- count(open, Borough)
# names(restaurant_borough)[names(restaurant_borough) == "n"] <- "count"
# rat_borough <- count(rat_311, Borough)
# borough <- c("Bronx", "Brooklyn", "Manhattan", "Queens", "Staten Island", "none")
# rat_borough <- cbind(rat_borough, borough) %>% filter(borough!= "none")
# rat_borough <- select(rat_borough, borough, n)
# names(rat_borough)[names(rat_borough) == "n"] <- "count"
# names(rat_borough)[names(rat_borough) == "borough"] <- "Borough"
# inspection_bc <- inspection %>% filter(GRADE == "B" | GRADE == "C")
# inspection_count_2020 <- count(inspection_bc, BORO)
# names(inspection_count_2020)[names(inspection_count_2020) == "n"] <- "count"
# names(inspection_count_2020)[names(inspection_count_2020) == "BORO"] <- "Borough"
# street_seating <- filter(open, Approved.for.Roadway.Seating == "yes")
# count_street <- count(street_seating, Borough)
# names(count_street)[names(count_street) == "n"] <- "count"
# sidewalk_seating <- filter(open, Approved.for.Sidewalk.Seating == "yes")
# count_sidewalk <- count(sidewalk_seating, Borough)
# names(count_sidewalk)[names(count_sidewalk) == "n"] <- "count"
# plot1 <- ggplot() + geom_bar(data=rat_borough, aes(x=Borough, y=count, fill=Borough), stat="identity") + ggtitle("2020 rodent reports") + ylab("Number of 311 calls\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot2 <- ggplot() + geom_bar(data=restaurant_borough, aes(x =Borough, y= count, fill =Borough), stat="identity") + ggtitle("Outdoor restaurants") + ylab("Applications approved\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), legend.position="none", axis.title.y=element_text(face="bold")) + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot3 <- ggplot() + geom_bar(data=inspection_count_2020, aes(x=Borough, y=count, fill=Borough), stat="identity") + ggtitle("Restaurants w/ B or C inspection scores") + ylab("Restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), legend.position="none", axis.title.y=element_text(face="bold")) + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot4 <- ggplot() + geom_bar(data=count_street, aes(x=Borough, y=count, fill=Borough), stat="identity") + ggtitle("Street dining") + ylab("Approved restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot5 <- ggplot() + geom_bar(data=count_sidewalk, aes(x=Borough, y=count, fill=Borough), stat="identity") + ggtitle("Sidewalk dining") + ylab("Approved restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot1a <- ggplotly(plot1, tooltip=c("x", "y"))
# plot2a <- ggplotly(plot2, tooltip=c("x", "y"))
# plot3a <- ggplotly(plot3, tooltip=c("x", "y"))
# plot4a <- ggplotly(plot4, tooltip=c("x", "y"))
# plot5a <- ggplotly(plot5, tooltip=c("x", "y"))
# plot1a
# plot2a
# plot3a
# plot4a
# plot5a
h6("Chart 9"),
plotlyOutput("brendan_chart1" ,width="50%"),
br(),
h6("Chart 10"),
plotlyOutput("brendan_chart2",width="50%"),
br(),
h6("Chart 11"),
plotlyOutput("brendan_chart3",width='50%'),
br(),
h6("Chart 12"),
plotlyOutput("brendan_chart4",width='50%'),
br(),
h6("Chart 13"),
plotlyOutput("brendan_chart5",width='50%'),
br(),
p(class = "padding", align = "left", "In all five figures, we can see that Manhattan is far above the rest of the boroughs
in restaurants approved for outdoor dining, in sidewalk and street dining, and B and C graded restaurants. However, it is
Brooklyn that is far above the rest of the boroughs in rodent reports in 2020. This suggests that perhaps another factor is
contributing to the rat problem in the Brooklyn borough. If restaurants were fully to blame for itβs rat problem, we would
expect to see it having high numbers of restaurants approved for outdoor street and sidewalk dining, a number comparable to
the borough of Manhattan. The first bar plot displays the number of rodent related 311 reports in the year 2020 by borough.
Brooklyn leads the way with well over 10,000 rodent related calls in the year, while the next closest borough, Manhattan, only
has about 8,000 rodent related calls in the year. In the bar plots related to restaurants, we see Manhattan leads the way across
the board. In the restaurants with outdoor dining, sidewalk and street dining, Manhattan has twice as many restaurants than any
other borough. Because of this vast difference in the number of restaurants in Manhattan compared to the other boroughs, we will
narrow our focus to Manhattan in the next visualization."),br(),br(),
),
),
br(),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings", href="https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9"), h6("Filtered to 311 calls from Jan 1 2020 to Dec 31 2020, of complaint type βrodentβ in Manhattan borough.")),
h5(a("Open Restaurants", href="https://data.cityofnewyork.us/Transportation/Open-Restaurant-Applications/pitm-atqc/data"),h6("Filtered to Manhattan borough.")),
),
div(class='padding',
h4("Visualization:"),
),
fluidRow(align = "center",
h3("Rat sightings in 2020 overlaid on restaurant locations in Manhattan"),
h6("Chart 14"),),br(),
fluidRow(
leafletOutput("brendan_map", height = 500, width = "100%"),
br(),
br(),
p(class = "padding", align = "left", "Based off exploratory analysis of the restaurants and rat reports across all boroughs, itβs clear
Manhattanβs restaurant industry may be most closely linked to the rat problem than in other boroughs.
For that reason, we have provided an interactive map visualization of the Manhattan borough
specifically. In the map, restaurants are plotted and viewers can see the location and name of the
restaurant, along with whether or not the restaurant is available for open street or sidewalk dining. Also
charted on the map are the location of rat sightings in the 2020 311 calls data set, the same data used
for previous visualizations. With no zoom, clusters of the rats are displayed in the visualization. After
zooming in further, those clusters break into the individual rat sighting locations of which they consist.
Rat sighting locations are represented by small rat icons on the map."),
p(class = "padding", align = "left", "This visualization allows viewers to identify densely rat populated locations on the map and relate that
information to the data provided for restaurants in the same locations. In real time, such a map would
be useful for avoidance of rat βhot spotsβ when choosing places to dine. It also allows users to explore
which restaurants have practices that may be contributing to the rat problem the most, with lots of rat
sightings nearby."),),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings", href="https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9"),
h6("Filtered to 311 calls from Jan 1 2020 to Dec 31 2020, of complaint type 'rodent' in Manhattan borough.
Calls related to noise, parking violations, and other non-emergency police related matters are also
excluded from this visualization.")),
),
div(class='padding',
h4("Visualization:"),
),
),
fluidRow(align = "center",
h3("Wordcloud of the descriptor variable of all NYC311 complaints in 2020"),
br(),
h6("Chart 15"),
img(src='Capture.PNG',width="50%"),
br(),
h6("Chart 16"),
img(src='Picture2.png',width="50%"),
br(),
h6("For reference:"),
img(src='rat.png',),
br(),
br(),
p(class = "padding", align = "left", "For required text analysis, we have again referred to the 2020 rodent related 311 reports, specifically on
the descriptor variable, where various notes are left on the nature or details of the complaint. Two word
clouds are presented. The first is a basic wordcloud created with the ggwordcloud package. Words
related to rat sightings are differentiated from others by color. Viewers can see in this wordcloud that
the descriptor variable does have lots of mentions of rodent related issues. The second wordcloud
presented is also from the ggwordcloud package, but with an added mask, intended to create a
wordcloud the shape of a rat. This visualization is slightly more visually appealing, but reveals the exact
same information to the reader. Rat sightings are often mentioned in the descriptor variable of the data
set."),br(),br()
),
# fluidRow(
# align = "center",
# plotOutput("brendan_wc1"),
# plotOutput("bendan_wc2"),
# ),
),
)
)
# code for generating the plots -----------------------------------------------------------------
server <- function(input, output, session) {
# prajwal's code for generating interactive map backend -------------------
# points <- eventReactive(input$recalc, {
# cbind(rat_sightings_sample$latitude, rat_sightings_sample$longitude)
# }, ignoreNULL = FALSE)
#
observe({
min_year <- input$year_input[1]
max_year <- input$year_input[2]
burough <- input$burough_input
case_status1 <- input$case_status
rat_sightings_sample <- rat_sightings[sample.int(nrow(rat_sightings), input$num_sample),]
#rat_sightings_sample <- rat_sightings
latitude_colnum <- grep('latitude', colnames(rat_sightings_sample))
longitude_colnum <- grep('longitude', colnames(rat_sightings_sample))
rat_sightings_sample <- rat_sightings_sample[complete.cases(rat_sightings_sample[,latitude_colnum:longitude_colnum]),]
rat_sightings_sample$year_created <- year(parse_date_time(rat_sightings_sample$Created.Date, '%m/%d/%y %I:%M:%S %p'))
#print('buroughh')
#print(burough)
#filter_rat_sightings <- rat_sightings_sample %>% filter(year_created >= min_year, year_created <= max_year, Borough %in% burough)
check_rows_of_filter <- nrow(rat_sightings_sample %>% filter(year_created >= min_year, year_created <= max_year,
Borough %in% burough,
Status %in% case_status1))
rat_sightings_buroughs2 <- as.character(unique(unlist(rat_sightings_sample$Borough)))
rat_sightings_case_status2 <- as.character(unique(unlist(rat_sightings_sample$Status)))
# print('buroughs 2')
# print(rat_sightings_buroughs2)
# print('case statuses 2')
# print(rat_sightings_case_status2)
if (check_rows_of_filter <= 0){
#updateSliderInput(session, "year_input", value = c(2010, 2021))
#updateSliderInput(session, "burough_input", value = rat_sightings_buroughs2)
#updateSliderInput(session, "case_status", value = rat_sightings_case_status2)
# filter_rat_sightings2 <- rat_sightings_sample
# reset("year_input")
# reset("burough_input")
# reset("burough_input")
# print('in the case of 0 rows, resetting to entire df')
leaflet() %>%
addProviderTiles(providers$Stamen.TonerLite,
options = providerTileOptions(noWrap = TRUE)
) %>%
setView(lng = -73.98928, lat = 40.75042, zoom = 10)
}
else{
filter_rat_sightings2 <- rat_sightings_sample %>% filter(year_created >= min_year, year_created <= max_year,
Borough %in% burough,
Status %in% case_status1)
filter_rat_sightings <- filter_rat_sightings2
getColor <- function(filter_rat_sightings, i) {
if(filter_rat_sightings$Status[i] == "Closed") {
"green"
}
else if(filter_rat_sightings$Status[i] == "In Progress") {
"lightblue"
}
else if(filter_rat_sightings$Status[i] == "Assigned") {
"orange"
}
else if(filter_rat_sightings$Status[i] == "Open") {
"purple"
}
else if(filter_rat_sightings$Status[i] == "Pending") {
"darkred"
}
else if(filter_rat_sightings$Status[i] == "Draft") {
"blue"
}}
markerColors <- rep(NA, nrow(filter_rat_sightings))
for (i in 1:nrow(filter_rat_sightings)){
markerColors[i] <- getColor(filter_rat_sightings, i)
}
icons <- awesomeIcons(
icon = 'ios-close',
iconColor = 'cadetblue',
library = 'ion',
markerColor = markerColors
)
output$map <- renderLeaflet({
leaflet(data = filter_rat_sightings) %>%
addProviderTiles(providers$Stamen.TonerLite,
options = providerTileOptions(noWrap = TRUE)
) %>%
setView(lng = -73.98928, lat = 40.75042, zoom = 10) %>%
addAwesomeMarkers( ~longitude, ~latitude, clusterOptions = markerClusterOptions() ,icon = icons,
popup = as.character(paste('Created date:', filter_rat_sightings$Created.Date,'<br>',
'Closed Date:', filter_rat_sightings$Closed.Date,'<br>',
#'Complaint type:',filter_rat_sightings$Complaint.Type,'<br>',
#'Descriptor:',filter_rat_sightings$Descriptor,'<br>',
'Address:',filter_rat_sightings$Incident.Address,'<br>',
'Status:', filter_rat_sightings$Status, '<br>',
'Location Type:', filter_rat_sightings$Location.Type))) %>%
addHeatmap( ~longitude, ~latitude, group = "heat",max=1, blur = 45, minOpacity = 0.8) %>% addLegend("topleft",
colors =c('green', "darkred", "lightblue", "orange"),
#"purple", "#56a0d1"
labels= c("Closed", "Pending", "In Progress","Assigned"),
# "Open", "Draft"
title= "Complaint status",
opacity = 1)
})
output$cityViz <- renderPlotly({
if (nrow(zipsInBounds()) == 0)
return(NULL)
tmp <- (zipsInBounds() %>% count(City))
tmp <- tmp[order(-tmp$n),]
tmp <- tmp[1:5,]
ggplotly(
ggplot(tmp, aes(x=City, y=n, fill = City)) + geom_bar(stat="identity") + ylab("Top 5 visible buroughs") + theme(legend.position = "none") + scale_color_brewer(palette="Dark2")+
theme(axis.title.x=element_blank(),
axis.ticks.x=element_blank())
)
})
output$locationViz <- renderPlotly({
if (nrow(zipsInBounds()) == 0)
return(NULL)
tmp <- (zipsInBounds() %>% count(Location.Type))
tmp <- tmp[order(-tmp$n),]
tmp <- tmp[1:5,]
ggplotly(tooltip = c("n"),
ggplot(tmp, aes(x=Location.Type, y=n)) + geom_bar(stat="identity", aes(fill = Location.Type)) + ggtitle('Visible location types') + ylab("Visible location types") +
theme_light()+
theme(axis.title.y=element_blank(),
#axis.text.y=element_blank(),
axis.title.x=element_blank(),
axis.text.x=element_blank(),
axis.ticks.x=element_blank()) + labs(fill = "Visible location types") + coord_flip() + theme(legend.position = "none")
)
})
output$yearViz <- renderPlotly({
if (nrow(zipsInBounds()) == 0)
return(NULL)
# total cases
created_date_sample <- data.table(zipsInBounds()$Created.Date)
created_date_sample$dates <- parse_date_time(created_date_sample$V1, '%m/%d/%y %I:%M:%S %p')
plot_created_year <- data.frame(table(year(date(created_date_sample$dates))))
for (i in 2010:2021){
if ((i %in% plot_created_year$Var1)==FALSE) {
#print(i)
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- c('Var1','Freq')
plot_created_year <- rbind(plot_created_year, tmp_df)
}
}
plot_created_year$Var1 <- as.numeric(as.character(plot_created_year$Var1))
names(plot_created_year)[names(plot_created_year) == "Var1"] <- "Year"
plot_created_year <- plot_created_year[order(plot_created_year$Year),]
plot_created_year <- filter(plot_created_year, Year >= min_year, Year <= max_year)
plot_created_year$case_status <- 'Total'
# closed cases
plot_closed<- filter(zipsInBounds(), Status == 'Closed')
flag_closed <- 0
if (nrow(plot_closed) == 0) {
flag_closed <- 1
}
if (flag_closed == 0){
plot_closed <- data.table(plot_closed$Created.Date)
plot_closed$dates <- parse_date_time(plot_closed$V1, '%m/%d/%y %I:%M:%S %p')
plot_closed_year <- data.frame(table(year(date(plot_closed$dates))))
#print(plot_closed_year)
for (i in 2010:2021){
if ((i %in% plot_closed_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_closed_year)
plot_closed_year <- rbind(plot_closed_year, tmp_df)
}
}
plot_closed_year$Var1 <- as.numeric(as.character(plot_closed_year$Var1))
names(plot_closed_year)[names(plot_closed_year) == "Var1"] <- "Year"
plot_closed_year <- plot_closed_year[order(plot_closed_year$Year),]
plot_closed_year <- filter(plot_closed_year, Year >= min_year, Year <= max_year)
plot_closed_year$case_status <- 'Closed'
}
# assigned cases
plot_assigned<- filter(zipsInBounds(), Status == 'Assigned')
flag_assigned <- 0
if (nrow(plot_assigned) == 0) {
flag_assigned <- 1
}
if (flag_assigned == 0){
plot_assigned <- data.table(plot_assigned$Created.Date)
plot_assigned$dates <- parse_date_time(plot_assigned$V1, '%m/%d/%y %I:%M:%S %p')
plot_assigned_year <- data.frame(table(year(date(plot_assigned$dates))))
#print(plot_assigned_year)
for (i in 2010:2021){
if ((i %in% plot_assigned_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_assigned_year)
plot_assigned_year <- rbind(plot_assigned_year, tmp_df)
}
}
plot_assigned_year$Var1 <- as.numeric(as.character(plot_assigned_year$Var1))
names(plot_assigned_year)[names(plot_assigned_year) == "Var1"] <- "Year"
plot_assigned_year <- plot_assigned_year[order(plot_assigned_year$Year),]
plot_assigned_year <- filter(plot_assigned_year, Year >= min_year, Year <= max_year)
plot_assigned_year$case_status <- 'Assigned'
#print('assigned or in progress')
#print(plot_assigned_year)
}
# assigned or in progress cases
plot_in_progress<- filter(zipsInBounds(), Status == 'In Progress')
flag_in_progress <- 0
if (nrow(plot_in_progress) == 0) {
flag_in_progress <- 1
}
if (flag_in_progress == 0){
plot_in_progress <- data.table(plot_in_progress$Created.Date)
plot_in_progress$dates <- parse_date_time(plot_in_progress$V1, '%m/%d/%y %I:%M:%S %p')
plot_in_progress_year <- data.frame(table(year(date(plot_in_progress$dates))))
#print(plot_in_progress_year)
for (i in 2010:2021){
if ((i %in% plot_in_progress_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_in_progress_year)
plot_in_progress_year <- rbind(plot_in_progress_year, tmp_df)
}
}
plot_in_progress_year$Var1 <- as.numeric(as.character(plot_in_progress_year$Var1))
names(plot_in_progress_year)[names(plot_in_progress_year) == "Var1"] <- "Year"
plot_in_progress_year <- plot_in_progress_year[order(plot_in_progress_year$Year),]
plot_in_progress_year <- filter(plot_in_progress_year, Year >= min_year, Year <= max_year)
plot_in_progress_year$case_status <- 'In Progress'
#print('assigned or in progress')
#print(plot_in_progress_year)
}
# open cases
plot_open<- filter(zipsInBounds(), Status == 'Open')
flag_open <- 0
if (nrow(plot_open) == 0) {
flag_open <- 1
}
if (flag_open == 0){
plot_open <- data.table(plot_open$Created.Date)
plot_open$dates <- parse_date_time(plot_open$V1, '%m/%d/%y %I:%M:%S %p')
plot_open_year <- data.frame(table(year(date(plot_open$dates))))
#print(plot_open_year)
for (i in 2010:2021){
if ((i %in% plot_open_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_open_year)
plot_open_year <- rbind(plot_open_year, tmp_df)
}
}
plot_open_year$Var1 <- as.numeric(as.character(plot_open_year$Var1))
names(plot_open_year)[names(plot_open_year) == "Var1"] <- "Year"
plot_open_year <- plot_open_year[order(plot_open_year$Year),]
plot_open_year <- filter(plot_open_year, Year >= min_year, Year <= max_year)
plot_open_year$case_status <- 'Open'
#print('open or pending')
#print(plot_open_year)
# print('created')
# print(plot_created_year)
# print('closed')
# print(plot_closed_year)
# print('combined')
}
# pending cases
plot_pending<- filter(zipsInBounds(), Status == 'Pending')
flag_pending <- 0
if (nrow(plot_pending) == 0) {
flag_pending <- 1
}
if (flag_pending == 0){
plot_pending <- data.table(plot_pending$Created.Date)
plot_pending$dates <- parse_date_time(plot_pending$V1, '%m/%d/%y %I:%M:%S %p')
plot_pending_year <- data.frame(table(year(date(plot_pending$dates))))
#print(plot_pending_year)
for (i in 2010:2021){
if ((i %in% plot_pending_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_pending_year)
plot_pending_year <- rbind(plot_pending_year, tmp_df)
}
}
plot_pending_year$Var1 <- as.numeric(as.character(plot_pending_year$Var1))
names(plot_pending_year)[names(plot_pending_year) == "Var1"] <- "Year"
plot_pending_year <- plot_pending_year[order(plot_pending_year$Year),]
plot_pending_year <- filter(plot_pending_year, Year >= min_year, Year <= max_year)
plot_pending_year$case_status <- 'Pending'
#print('open or pending')
#print(plot_pending_year)
# print('created')
# print(plot_created_year)
# print('closed')
# print(plot_closed_year)
# print('combined')
}
#plot_this <- plot_created_year
plot_this <- data.frame(matrix(ncol = 3, nrow = 0))
colnames(plot_this) <- c("Year",'Freq','case_status')
rownames(plot_created_year)
# print('flag open')
# print(flag_open)
# print('flag pending')
# print(flag_pending)
# print('flag assigned')
# print(flag_assigned)
# print('flag in progress')
# print(flag_in_progress)
# print('flag closed')
# print(flag_closed)
if (flag_open == 0){
plot_this <- rbind(plot_this, plot_open_year)
}
if (flag_pending == 0){
plot_this <- rbind(plot_this, plot_pending_year)
}
if (flag_assigned == 0) {
plot_this <- rbind(plot_this, plot_assigned_year)
}
if (flag_in_progress == 0) {
plot_this <- rbind(plot_this, plot_in_progress_year)
}
if (flag_closed == 0){
plot_this <- rbind(plot_this, plot_closed_year)
}
#print(plot_this)
# p_years <- ggplotly(
# ggplot(data=plot_created_year, aes(x=Year, y=Freq)) + geom_path(stat="identity") + ylab('Rat sightings') + geom_point()+
# theme(axis.title.x=element_blank()) + scale_x_continuous(breaks=seq(min_year, max_year, 1))
# )
colors_for_case_status <- rep(NA, nrow(plot_this))
for (i in 1:nrow(plot_this)){
if (plot_this$case_status[i] == "Closed"){
colors_for_case_status[i] <- "green"
}
else if (plot_this$case_status[i] == "Pending"){
colors_for_case_status[i] <- "darkred"
}
else if (plot_this$case_status[i] == "Assigned"){
colors_for_case_status[i] <- "orange"
}
else if (plot_this$case_status[i] == "In Progress"){
colors_for_case_status[i] <- "cadetblue"
}
else if (plot_this$case_status[i] == "Draft"){
colors_for_case_status[i] <- "blue"
}
else if (plot_this$case_status[i] == "Open"){
colors_for_case_status[i] <- "purple"
}
}
p_years <- ggplotly(
ggplot(data=plot_this, aes(x=Year, y=Freq)) + geom_line(aes(color=case_status)) + geom_point(aes(colour=case_status))
#+ scale_colour_manual(name = 'Case status',values =c('green'='green','cadetblue' = 'cadetblue', 'orange'='orange', 'darkred'='darkred'), labels = c("closed","in progress", "assigned",'pending'))
+ ggtitle('Complaint status trend') + scale_x_continuous(breaks=seq(min_year, max_year, 1)) + theme_light()
+ theme(axis.title.y=element_blank(),
#axis.text.y=element_blank(),
axis.title.x=element_blank(),)
#axis.text.x=element_blank())
+ theme(legend.title = element_blank())
#+ theme(legend.position="left")
#+ labs(color='Status')
+
theme(axis.title.x=element_blank())
)
})
zipsInBounds <- reactive({
if (is.null(input$map_bounds))
return(zipdata[FALSE,])
bounds <- input$map_bounds
#print(bounds)
latRng <- range(bounds$north, bounds$south)
lngRng <- range(bounds$east, bounds$west)
#print(latRng)
subset(filter_rat_sightings,
latitude >= latRng[1] & latitude <= latRng[2] &
longitude >= lngRng[1] & longitude <= lngRng[2])
})
}
#filter_rat_sightings <- filter_rat_sightings[,burough]
# if (nrow(event_data("plotly_selecting"))>0){
# filter_rat_sightings <- filter_rat_sightings %>% filter(year_created %in% event_data("plotly_selecting")$Var1)
# }
#print('reached here')
#print(filter_rat_sightings)
#print('end reached here')
})
# PRATISHTA'S VISUALIZATIONS -------------------------------------------------
output$pratishta1 <- renderPlotly({
p <- ggplot(rat_ton_date, aes(x=Created.Date, y=REFUSETONSCOLLECTED)) +
geom_line(aes(color = Borough)) +
geom_point(aes(color = Borough)) +
xlab("Date by Months") +
ylab("Weight of Waste (Tons)") + theme_light()
p
})
output$pratishta2 <- renderPlotly({
p <- ggplot(rat_ton_date, aes(x=Created.Date, y=n)) +
geom_line(aes(color = Borough)) +
geom_point(aes(color = Borough)) +
xlab("Date by Months") +
ylab("Number of rat sightings") + theme_light()
p
})
output$pratishta3 <- renderPlotly({
p <- ggplot(rat_ton_date, aes(x=Created.Date, y=rate)) +
geom_line(aes(color = Borough)) +
geom_point(aes(color = Borough)) +
xlab("Date by Months") +
ylab("Rate of rats per kiloton of waste") + theme_light()
p
})
output$pratishta4 <- renderPlotly({
ton +
xlab("Boroughs") +
ylab("Weight of Waste (Tons)") + theme_light()
})
output$pratishta5 <- renderPlotly({
rat +
xlab("Boroughs") +
ylab("Number of Rat Sightings") + theme_light()
})
output$pratishta7 <- renderTmap({
## ------------------------------------------------------------------------
tm_shape(nyc_sp) +
tm_fill("n", title = "Rat Sightings in Community Districts")+tm_view(set.view = c(-73.98928, 40.70042,10))
})
output$pratishta8 <- renderTmap({
## ------------------------------------------------------------------------
tm_shape(nyc_sp) +
tm_fill("total_ton", title = "Tones of Waste and Rat Sightings by DSNY Districts")+tm_view(set.view = c(-73.98928, 40.70042,10))
})
output$pratishta9 <- renderPlot({
## ------------------------------------------------------------------------
legend_creator = function(col.regions, xlab, ylab, nbins){
bilegend = levelplot(matrix(1:(nbins * nbins), nrow = nbins),
axes = FALSE, col.regions = col.regions,
xlab = xlab, ylab = ylab,
cuts = 8, colorkey = FALSE, scales = list(draw = 0))
bilegend
}
add_new_var = function(x, var1, var2, nbins, style = "quantile"){
class1 = suppressWarnings(findCols(classIntervals(c(x[[var1]]),
n = nbins,
style = style)))
class2 = suppressWarnings(findCols(classIntervals(c(x[[var2]]),
n = nbins,
style = style)))
x$new_class = class1 + nbins * (class2 - 1)
return(x)
}
## ------------------------------------------------------------------------
nyc_cd <- nyc_sp
## ------------------------------------------------------------------------
nyc_cd = add_new_var(nyc_cd,
var1 = "n",
var2 = "total_ton",
nbins = 3)
## ------------------------------------------------------------------------
bilegend = legend_creator(stevens.pinkblue(n = 9),
xlab = "rats",
ylab = "total tonnes",
nbins = 3)
vp = viewport(x = 0.25, y = 0.25, width = 0.25, height = 0.25)
pushViewport(vp)
#print(bilegend, newpage = FALSE)
bimap = tm_shape(nyc_cd) +
tm_fill("new_class", style = "cat", palette = stevens.pinkblue(n = 9), legend.show = FALSE) +
tm_layout(legend.show = FALSE)
grid.newpage()
vp = viewport(x = 0.37, y = 0.75, width = 0.25, height = 0.25)
print(bimap, vp = viewport())
pushViewport(vp)
print(bilegend, newpage = FALSE, vp = vp)
})
# END OF PRATISHTA'S VISUALIZATIONS ------------------------------------------
# brendan's viz -----------------------------------------------------------
output$brendan_chart1 <- renderPlotly({
plot1 <-ggplot() + geom_bar(data=rat_borough, aes(x=Borough, y=count), stat="identity", fill=rainbow(n=5)) + ggtitle("2020 rodent reports") + ylab("Number of 311 calls\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot1
})
output$brendan_chart2 <- renderPlotly({
plot2 <- ggplot() + geom_bar(data=restaurant_borough, aes(x =Borough, y= count), stat="identity", fill=rainbow(n=5)) + ggtitle("Outdoor restaurants") + ylab("Applications approved\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), legend.position="none", axis.title.y=element_text(face="bold")) + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot2
})
output$brendan_chart3 <- renderPlotly({
plot3 <- ggplot() + geom_bar(data=inspection_count_2020, aes(x=Borough, y=count), stat="identity", fill=rainbow(n=5)) + ggtitle("Restaurants w/ B or C inspection scores") + ylab("Restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), legend.position="none", axis.title.y=element_text(face="bold")) + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot3
})
output$brendan_chart4 <- renderPlotly({
plot4 <- ggplot() + geom_bar(data=count_street, aes(x=Borough, y=count), stat="identity", fill=rainbow(n=5)) + ggtitle("Street dining") + ylab("Approved restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot4
})
output$brendan_chart5 <- renderPlotly({
plot5 <- ggplot() + geom_bar(data=count_sidewalk, aes(x=Borough, y=count), stat="identity", fill=rainbow(n=5)) + ggtitle("Sidewalk dining") + ylab("Approved restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot5
})
output$brendan_map <- renderLeaflet({
interactive_map <- leaflet() %>% addProviderTiles(providers$Stamen.TonerLite,
options = providerTileOptions(noWrap = TRUE)) %>% addCircleMarkers(data=manhattan_open, lng=~Longitude, lat=~Latitude, radius=1, color= "red", fillOpacity=1, popup=~paste('Restuarant:', manhattan_open$Restaurant.Name, '<br>', 'Street Dining?', manhattan_open$Approved.for.Roadway.Seating, '<br>', 'Sidewalk Dining?', manhattan_open$Approved.for.Sidewalk.Seating, '<br>')) %>% addMarkers(data=manhattan_311,lng=~manhattan_311.Longitude, lat=~manhattan_311.Latitude, icon=icon, popup=~paste('Address:',manhattan_311$manhattan_311.Incident.Address,'<br>', 'Call Date:', manhattan_311$manhattan_311.Created.Date,'<br>','Descriptor:', manhattan_311$manhattan_311.Descriptor,'<br>'), clusterOptions= markerClusterOptions(disableClusteringAtZoom=16)) %>%
setView(lng = -73.98928, lat = 40.77042, zoom = 12)
})
# output$brendan_wc1 <- renderPlot({
# (wordcloud1 <- ggplot(descriptors3, aes(label=word, size=n, color=col)) + geom_text_wordcloud_area(mask=img, rm_outside = TRUE) + scale_size_area(max_size=5) + theme_classic())
#
# })
#
# output$brendan_wc2 <- renderPlot({
# (wordcloud2 <- ggplot(descriptors3, aes(label=word, size=n, color=col)) + geom_text_wordcloud_area() + scale_size_area())
#
# })
}
shinyApp(ui = ui, server = server)
| /group_n/app.R | no_license | pratishta/Group_N_NYCdata | R | false | false | 66,717 | r | #install.packages('rsconnect')
#library(rsconnect)
#install.packages("shinyjs")
options(scipen=999)
# imports -----------------------------------------------------------------
library(shiny)
library(shinyjs)
library(leaflet)
library(RColorBrewer)
library(scales)
library(lattice)
library(dplyr)
library(ggplot2)
library(plotly)
library(data.table)
library(lubridate)
library(leaflet.extras)
library(magrittr) # chain operators, e.g. to "pipe" a value forward
#library(plyr)
library(tidyverse)
library(DT)
library(knitr)
library(maps)
library(rgdal)
library(ggmap)
library(tmap)
library(sp)
library(tmap)
library(sf)
library(stars)
library(spData)
library(classInt)
library(lattice)
library(grid)
library(pals)
# prajwal's data fetch ----------------------------------------------------------
# Download the data from https://data.cityofnewyork.us/api/views/3q43-55fe/rows.csv?accessType=DOWNLOAD
# Alternate link: https://data.cityofnewyork.us/Social-Services/Rat-Sightings/3q43-55fe, click Export -> CSV
rat_sightings <- read.csv("https://data.cityofnewyork.us/api/views/3q43-55fe/rows.csv?accessType=DOWNLOAD")
#rat_sightings <- read.csv("data/Rat_Sightings.csv")
rat_sightings <- rat_sightings %>% filter(!(Status=='Open'))
rat_sightings <- rat_sightings %>% filter(!(Status=='Draft'))
rat_sightings <- rat_sightings %>% filter(!(Borough=='Unspecified'))
rat_sightings$latitude <- rat_sightings$Latitude
rat_sightings$longitude <- rat_sightings$Longitude
#set.seed(100)
#c("BROOKLYN", "QUEENS","STATEN ISLAND")
#rat_sightings_buroughs <- c("BROOKLYN", "QUEENS","STATEN ISLAND")
# PRATISHTA'S DATA FETCH -------------------------------------------------------
# read in the main csv file
rat_data<-read.csv("data/rat_data.csv")
rat_data <- rat_data %>%
mutate(Borough = str_to_title(rat_data$Borough))
tonnage_data<-read.csv("data/dsny_boro_tonnage.csv", stringsAsFactors = FALSE)
ton_date <- tonnage_data %>%
mutate(MONTH = paste(MONTH, " / 01")) %>%
mutate(MONTH = as.Date(MONTH, format = '%Y / %m / %d')) %>%
filter(MONTH > as.Date('2020-01-01', '%Y-%m-%d'), MONTH < as.Date('2021-03-01', '%Y-%m-%d')) %>%
arrange(desc(MONTH))
rat_date <- rat_data %>%
mutate(Created.Date = as.Date(Created.Date, "%m/%d/%Y")) %>%
mutate(Created.Date = as.character(Created.Date)) %>%
mutate(Created.Date = substr(Created.Date, 1, 8)) %>%
mutate(Created.Date = paste(Created.Date, '01')) %>%
mutate(Created.Date = as.Date(Created.Date, "%Y-%m-%d")) %>%
group_by(Created.Date, Borough) %>%
tally() %>%
filter(Created.Date > as.Date('2020-01-01', '%Y-%m-%d'), Created.Date < as.Date('2021-03-01', '%Y-%m-%d')) %>%
arrange(desc(Created.Date))
rat_ton_date <- merge(rat_date, ton_date, by.x = c("Created.Date", "Borough"), by.y = c("MONTH", "BOROUGH")) %>%
mutate(rate = n / (REFUSETONSCOLLECTED / 100))
# END OF PRATISHTA'S DATA FETCH ------------------------------------------------
# PRATISHTA'S CODE -------------------------------------------------------------
# community district conversion functions
convertBoroCDToDistrict <- function(borocd) {
sapply(borocd, function(borocd) {
boro_ch = as.character(borocd)
boro_n = substr(boro_ch, 1, 1)
cd_n = substr(boro_ch, 2, 3)
boro = case_when (boro_n == '1' ~ 'MN',
boro_n == '2' ~ 'BX',
boro_n == '3' ~ 'BK',
boro_n == '4' ~ 'QW',
boro_n == '5' ~ 'SI'
)
ans <- paste(boro, cd_n, sep="")
return (ans)
})
}
convertToShpDistrict <- function(com_district) {
sapply(com_district, function(com_district) {
split = strsplit(com_district, " ")
boro = case_when (str_to_lower(split[[1]][2]) == 'brooklyn' ~ 'BK',
str_to_lower(split[[1]][2]) == 'manhattan' ~ 'MN',
str_to_lower(split[[1]][2]) == 'queens' ~ 'QW',
str_to_lower(split[[1]][2]) == 'staten' ~ 'SI',
str_to_lower(split[[1]][2]) == 'bronx' ~ 'BX'
);
ans <- paste(boro, split[[1]][1], sep="")
return (ans)
})
}
# reading in data and modify string format of community district column
full_tonnage <-read.csv("sanitation_data/dsny_full_tonnage.csv", stringsAsFactors = FALSE)
full_tonnage <- full_tonnage %>%
mutate(district = paste(full_tonnage$COMMUNITYDISTRICT, str_to_upper(full_tonnage$BOROUGH)))
district = paste(full_tonnage$COMMUNITYDISTRICT, str_to_upper(full_tonnage$BOROUGH))
# creating data to be mapped
ton_map <- full_tonnage %>%
mutate(community_district = convertToShpDistrict(district)) %>%
group_by(community_district) %>%
summarise(total_ton = sum(REFUSETONSCOLLECTED))
ton_map
community_district <- paste(rat_data$Community.Board, str_to_upper(rat_data$Community.Board))
rat_map <- rat_data %>%
mutate(community_district = convertToShpDistrict(community_district)) %>%
group_by(community_district) %>%
tally()
rat_map
rat_borough <- rat_data %>%
group_by(Borough) %>%
tally()
rat_borough
ton_boro <- tonnage_data %>%
group_by(BOROUGH) %>%
summarise(total_ton = sum(REFUSETONSCOLLECTED))
ton_boro
rat_ton <- left_join(rat_borough, ton_boro, by = c("Borough" = "BOROUGH"))
rat <- ggplot(rat_ton, aes(y=n, x=Borough, fill = Borough)) +
geom_bar(position="dodge", stat="identity")
ton <- ggplot(rat_ton, aes(y=total_ton, x=Borough, fill = Borough)) +
geom_bar(position="dodge", stat="identity")
# map data
nyc <- readOGR("sanitation_data/CommunityDistricts/.", "geo_export_d81daad1-2b49-44c3-81d4-72436a58def3")
nyc_sp <- spTransform(nyc, CRS("+proj=longlat +datum=WGS84"))
nyc_sp@data <- nyc_sp@data %>%
mutate(community_district = convertBoroCDToDistrict(boro_cd))
nyc_sp@data
nyc_sp@data <- left_join(nyc_sp@data, rat_map)
nyc_sp@data
nyc_sp@data <- left_join(nyc_sp@data, ton_map)
nyc_sp@data
# END OF PRATISHTA'S CODE-------------------------------------------------------
# shiny code
# brendan's imports and code -------------------------------------------------------
library(ggplot2)
library(ggthemes)
library(gridExtra)
library(dplyr)
library(readr)
library(leaflet)
library(leaflet.extras)
library(magrittr)
library(dplyr)
library(tidyr)
library(wordcloud)
library(png)
library(ggwordcloud)
library(tidytext)
library(readr)
library(png)
#setwd("~/Bayesian")
open <- read.csv("data/Open_Restaurant_Applications.csv")
inspection <- read.csv("data/DOHMH_New_York_City_Restaurant_Inspection_Results.csv")
rat_311 <- read.csv("data/311_Service_Requests_from_2010_to_Present.csv")
restaurant_borough <- count(open, Borough)
names(restaurant_borough)[names(restaurant_borough) == "n"] <- "count"
rat_borough <- count(rat_311, Borough)
borough <- c("Bronx", "Brooklyn", "Manhattan", "Queens", "Staten Island", "none")
rat_borough <- cbind(rat_borough, borough) %>% filter(borough!= "none")
rat_borough <- select(rat_borough, borough, n)
names(rat_borough)[names(rat_borough) == "n"] <- "count"
names(rat_borough)[names(rat_borough) == "borough"] <- "Borough"
inspection_bc <- inspection %>% filter(GRADE == "B" | GRADE == "C")
inspection_count_2020 <- count(inspection_bc, BORO)
names(inspection_count_2020)[names(inspection_count_2020) == "n"] <- "count"
names(inspection_count_2020)[names(inspection_count_2020) == "BORO"] <- "Borough"
street_seating <- filter(open, Approved.for.Roadway.Seating == "yes")
count_street <- count(street_seating, Borough)
names(count_street)[names(count_street) == "n"] <- "count"
sidewalk_seating <- filter(open, Approved.for.Sidewalk.Seating == "yes")
count_sidewalk <- count(sidewalk_seating, Borough)
names(count_sidewalk)[names(count_sidewalk) == "n"] <- "count"
manhattan_311 <- filter(rat_311, Borough == "MANHATTAN")
#manhattan_311 <- read.csv("manhattan311.csv")
manhattan_open <- read.csv("data/manhattan open restaurants.csv")
manhattan_311 <- filter(manhattan_311, manhattan_311$Complaint.Type=="Rodent")
manhattan_311 <- data.frame(manhattan_311$Latitude, manhattan_311$Longitude, manhattan_311$Incident.Address, manhattan_311$Created.Date, manhattan_311$Descriptor)
icon <- makeIcon(iconUrl= "https://cdn3.iconfinder.com/data/icons/farm-animals/128/mouse-512.png", iconWidth=25, iconHeight = 20)
#manhattan_311b <- read.csv("manhattan311.csv")
#manhattan_311b <- read.csv("https://nycopendata.socrata.com/api/views/erm2-nwe9/rows.csv?accessType=DOWNLOAD")
# code for making wordcloud (not used in live version for processing power reasons) -----------------------------------------------
# manhattan_311b <- filter(rat_311, Borough == "MANHATTAN")
# nrow(manhattan_311b)
#
# manhattan_311b <- manhattan_311b %>% filter(Complaint.Type != "Noise = Street/Sidewalk" & Complaint.Type != "Noise - Residential" & Complaint.Type != "HEAT/HOT WATER" & Complaint.Type != "Illegal Parking" & Complaint.Type != "Non-Emergency Police Matter" & Complaint.Type != "Noise" & Complaint.Type != "Noise - Vehicle" & Complaint.Type != " Noise - Commercial")
# descriptors <- manhattan_311b %>% select(Descriptor)
# descriptors_fix <- as.character(descriptors$Descriptor)
# text_df <- tibble(line = 1:length(descriptors_fix), text = descriptors_fix)
# descriptors <- text_df %>% unnest_tokens(word, text)
#
# descriptors <- count(descriptors, word)
# descriptors2 <- filter(descriptors, n > 2000)
# col <- c(ifelse(descriptors2$word == "pests" | descriptors2$word == "rat" | descriptors2$word == "sighting" | descriptors2$word == "rodents", "red", "black"))
# descriptors3 <- cbind(descriptors2, col)
# descriptors3 <- filter(descriptors3, word != "n" & word != "a" & word != "not" & word != "business" & word != "no" & word!= "compliance" & word != "or" & word != "in" & word != "of" & word!= "to" & word!= "non" & word!= "on" & word != "has" & word!= "for")
# #setwd("~/Bayesian")
# img <- readPNG("data/rat.png")
# #img <- icon
# descriptors3 <- descriptors3 %>% filter(word != "loud" & word!= "music" & word != "party")
# set.seed(14)
# (wordcloud1 <- ggplot(descriptors3, aes(label=word, size=n, color=col)) + geom_text_wordcloud_area(mask=img, rm_outside = TRUE) + scale_size_area(max_size=5) + theme_classic())
# (wordcloud2 <- ggplot(descriptors3, aes(label=word, size=n, color=col)) + geom_text_wordcloud_area() + scale_size_area())
# user interface for setting layout of plots ----------------------------------------------------------
# sliders and interactive map ---------------------------------------------
rat_sightings_buroughs <- as.character(unique(unlist(rat_sightings$Borough)))
rat_sightings_case_status <- as.character(unique(unlist(rat_sightings$Status)))
ui <- fluidPage(
# header description ------------------------------------------------------
tags$head(
# Note the wrapping of the string in HTML()
tags$style(HTML("
.row {
margin-left: 0;
margin-right:0;
}"))
),
fluidRow(align = "center",
h1("Rats and NYC: Exploratory Visualization"),
strong("Data Visualization (QMSS - G5063) Final Project, Spring 2021"),
br(),
em("Group N: Brendan Mapes, Prajwal Seth, and Pratishta Yerakala"),
h3(a("Link to code and process book", href="https://github.com/QMSS-G5063-2021/Group_N_NYCdata")),
br(),br(),br(),
p("In this project, we will explore in detail New York City's rat problem.
New York has always dealt with a relatively high population of rats, as do many
other large metropolitan areas. However, since the beginning of the COVID-19 pandemic
rat sightings have been on the rise. Rats are being seen more often, during new times of day,
and are acting more aggressive. Through the following visualizations we hope to find some explanation
for this recent uptick in rat sightings. The way restaurants and residents handle their trash plays
a large role in the survival and behavior of rats in the city. So through exploration of city
sanitation data, restaurant registration data, and 311 calls in the city, we hope to find some
potential explanations as to why New York's rat problem has gotten so bad."),),
br(),br(),br(),
# prajwal's description ---------------------------------------------------
# fluidRow(
# align = "center",
# headerPanel("Hello 1!"),
# p("p creates a paragraph of text."),
# p("A new p() command starts a new paragraph. Supply a style attribute to change the format of the entire paragraph.", style = "font-family: 'times'; font-si16pt"),
# strong("strong() makes bold text."),
# em("em() creates italicized (i.e, emphasized) text."),
# br(),
# code("code displays your text similar to computer code"),
# div("div creates segments of text with a similar style. This division of text is all blue because I passed the argument 'style = color:blue' to div", style = "color:blue"),
# br(),
# p("span does the same thing as div, but it works with",
# span("groups of words", style = "color:blue"),
# "that appear inside a paragraph."),
# ),
fluidRow(
align = "center",
style='margin-left:0px; margin-right: 0px;',
h2("Interactive map of rodent complaints in NYC since 2010"),
h3("Prajwal Seth"),
br(),
),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings (automatically updated daily)", href="https://data.cityofnewyork.us/Social-Services/Rat-Sightings/3q43-55fe"),
br(),
h4("Background:"),
p("In this section, I have visualized rodent complaints from 2010 till today submitted to the NYC311 portal.
Feel free to play around with the provided filters for number of samples,
years, boroughs, and case status (however, due to processing constraints on shinyapps.io,
the website will crash if you set the number of samples too high). The map on the left will dynamically update as you
change the filters (or will not update if there is no data to display after the filters are applied).
The plot for the trend in complaint status also updates according to the
rat complaints that are visible in the map. Upon zooming into the map, you will see that the color of the
marker for each complaint is set according to the complaint status (refer to the legend of the map).
Also provided is a tooltip displaying the complaint's created date, closed date, address, status, and
location type. There is a layer of heat added to the map, with the intensity of the heat
being calculated based on the number of rat sightings in the area."),
),
),
# sliders and map etc -----------------------------------------------------
fluidRow(
sidebarLayout(position = "right",
sidebarPanel(width = 6,
sliderInput("num_sample", label = h4("Select number of random samples"), min = 1,
max = nrow(rat_sightings), value = 1000, step = 1000),
sliderInput("year_input", label = h4("Select years"), min = 2010,
max = 2021, value = c(2010, 2021), step = 1, format = "####"),
#selected = rat_sightings_buroughs[1:length(multiInput)])
#selected = rat_sightings_buroughs,
selectizeInput("burough_input", label=h4("Select boroughs"), choices =rat_sightings_buroughs, multiple = TRUE, selected = rat_sightings_buroughs),
selectizeInput("case_status", label=h4("Select status"), choices =rat_sightings_case_status, multiple = TRUE, selected = rat_sightings_case_status),
#plotlyOutput("cityViz", height = 300),
plotlyOutput("yearViz", height = 250),
#plotlyOutput("locationViz", height = 220),
#plotlyOutput("locationViz", height = 300),
),
mainPanel(width = 6, style='margin-top:40px;',
leafletOutput("map", height = 825),
),
),
# PRATISHTA'S WRITEUP --------------------------------------------------------
fluidRow(
align = "center",
style='margin-left:0px; margin-right: 0px;',
h2("Rat Sightings and Sanitation Waste by Borough"),
h3("Pratishta Yerakala"),
br(),
),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings", href="https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9"), h6("filtered for rodent sightings between Feb 2020 and Feb 2021.")),
h5(a("DSNY Monthly Tonnage", href="https://data.cityofnewyork.us/City-Government/DSNY-Monthly-Tonnage-Data/ebb7-mvp5"), h6("filtered for months between Feb 2020 and Feb 2021.")),
h5(a("NYC Community Districts Shapefile", href="https://data.cityofnewyork.us/City-Government/Community-Districts/yfnk-k7r4")),
),
div(class='padding',
h4("Background:"),
p("The large rodent population in New York City is no secret. Rats have been
associated with the city for a long time whether it's from the famous", a('pizza rat', href='https://knowyourmeme.com/memes/pizza-rat'),
"or to the rising concern from residents who have noticed changes since the COVID-19 pandemic.
Many businesses and normally occurring procedures have been slowed down or
halted completely. One such example in particular with the Department of
Sanitation of NY (DSNY) where limited resources and budget cuts since the
pandemic have caused an", a("increased amount of litter and waste production", href="https://patch.com/new-york/new-york-city/city-state-leaders-decry-sanitation-setback-trash-piles"),
"."),
)
),
fluidRow(
align = "center",
div(class='padding',
h4(align = "left", "Visualizations:"),
),
# descriptive charts
h3("Total Number of Rat Sightings (2020-2021)"),
h6("Chart 1"),
plotlyOutput("pratishta5", width = "50%"),
br(),
h3("Total Waste Produced (2020-2021)"),
h6("Chart 2"),
plotlyOutput("pratishta4", width = "50%"),
br(),
p(class = "padding", align = "left", "We can see in Chart 1 that Brooklyn produces the most tons of waste followed by
Queens, and then by Bronx and Manhattan. Staten Island is last with the
least amount of waste. Chart 2 shows the number of rat sightings per
borough and naturally, we have Brooklyn at the top with around 6,000
sightings. But Instead of Queens, Manhattan follows with most rat s
ightings. Then Queens and Bronx. From here it seems that Staten Island and
Bronx are boroughs that have some what proportional sightings to waste
produced. However, though Queens produces a lot of waste, it does not have
nearly the same rate of rat sightings. Conversely, Manhattan doesn't quite
produce the same amount of waste as Queens but seems to have far more rat
sightings. Brooklyin is consistenly infested."),
# time series charts
h3("Waste per Month (2020-2021)"),
h6("Chart 3"),
plotlyOutput("pratishta1", width = "70%"),
br(),
h3("Rat Sightings per Month (2020-2021)"),
h6("Chart 4"),
plotlyOutput("pratishta2", width = "70%"),
br(),
h3("Rate of Rats per Waste Ton per Month"),
h6("Chart 5"),
plotlyOutput("pratishta3", width = "70%"),
br(),
p(class = "padding", align = "left", "Charts 3 to 5 show a time series line graphs. Chart 3 shows the tons of
waste per month generated by each borough. It seems that around March and
April of 2020 there was a change in the trend. Though the waste was
rising, it flattened out - or even fell like with Manhattan for example
- between March and April of 2020. But after april there was a more mellow
rise and then gentle decline closer to Fall of 2020 and early 2021. This
exploratory chart is good to possibly check out Manhattan and why the
waste production went down. Perhaps for restaurants closing?"),
p(class = "padding", align = "left", "Chart 4 shows that all boroughs saw an increase in rat sightings,
especially Brooklyn. It did seem to peak around summer of 2020 and decline
again to almost normal rates. These sightings might be due to the
sanitation departments' limits as mentioned earlier."),
p(class = "padding", align = "left", "Chart 5 looks at the 'rate' at which a rat sighting is possible for every
kilo-ton of waste produced in each borough per month. It seems that these
rates also follow a similar path of an increase around April 2020 and then
a peak in summer of 2020 and then a mellow decline. However Manhattan's
rate shot up for early 2021 with around 0.75 (sightings per ton of waste)
to 1.5 (sightings per ton of waste). Perhaps this could be due to frequent
testings, vaccinations, and re-opening of restaurants (producing more
waste)?"),
# maps
h3("Number of Rat Sightings per Month"),
h6("Chart 6"),
),
fluidRow(
tmapOutput("pratishta7", width = "100%"),
br(),
),
fluidRow(align = "center",
h3("Waste Produced in Tons"),
h6("Chart 7"),
),
fluidRow(
tmapOutput("pratishta8", width = "100%"),
br(),
),
fluidRow(
align = "center",
h3("Rat Sightings and Waste Produced By Community District"),
h6("Chart 8"),
plotOutput("pratishta9", width = "60%"),
br(),
p(class = "padding", align = "left", "Here we show choropleth maps of NYC by
community district. Community district as the geographical feature here is
because DSNY also records their information with that feature. It seems
that there is one community districte that's seeing a severe rise in rat
sightings (BK03). It may be worth it to take a closer look at that
particular district if analysis or further studie are done."),
p(class = "padding", align = "left", "As expected it looks like Queens
produces a massive amount of waste and the particular districts are
highlighted in deep red (100,000 - 120,000 tons category). And though
Brooklyn also produces a lot of waste it seems to be spread out amongst
the community districts."),
p(class = "padding", align = "left", "Chart 8 depicts a bivariate choropleth
map of the community districts. This make-shift rendering of a map using
the tm_map package is from", a("this currently open issue from the tmap GitHub
repsitory", href="https://github.com/mtennekes/tmap/issues/183#issuecomment-670554921"),
". As the districts become more blue in color, the more right sightings
that have been reported in that district. The more purple a district, the
more amount of waste produced there. As the color goes to indigo, there's
a high rat sighting and a high waste production (a high high category if
further spatial dependence analysis was conducted). Light grey represents
few rat sightings and little waste production (a 'low low' again for further
spatial research). This map demonstrates that hough there are high rat
sightings in downtown Manhattan and parts of Brooklyn, it's not
necessarily tied to waste production. In the same way, the outer boroughs
have a huge waste production but not many rat sightings. But there are
some districts (indigo) that do exhibit both features in high amounts.
This exploratory data visulaization provides the insight to further look
in those districts.")
),
# Brendan's writeup -------------------------------------------------
fluidRow(
align = "center",
h2("Rat Sightings and Restaurants by Borough"),
h3("Brendan Mapes"),
br(),
),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings", href="https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9"), h6("Filtered to 311 calls from Jan 1 2020 to Dec 31 2020, of complaint type βrodentβ.")),
h5(a("Open Restaurants", href="https://data.cityofnewyork.us/Transportation/Open-Restaurant-Applications/pitm-atqc/data"),),
),
div(class='padding',
h4("Background:"),
p("It is well documented that the COVID-19 pandemic has led to a rise in rat sightings throughout New York
City. It is also well known that the pandemic has been especially hard on the restaurant industry. Hit
especially hard by stay-at-home orders and social distancing mandates, restaurants have been forced to
innovate their operations. Now more than ever, restaurants are serving customers outside in the
streets. We suspect this change in the way restaurants do their business may be contributing to the
increase in rat sightings. We explore this possibility a bit further in the next few visualizations."),
),
div(class='padding',
h4("Visualizations:"),
),
),
# descriptive charts
# fluidRow(align = "center",
# br(),
# h3("NYC311 rodent complaints in 2020 and number of restaurants by type"),
# h6("Chart 9"),br(),
# plotOutput("brendan_chart1", width = "80%"),
# plotOutput("brendan_chart2", width = "80%"),
# br(),
# br(),
# p(class = "padding", align = "left", "In all four figures, we can see that Manhattan is far above the rest of the boroughs in restaurants
# approved for outdoor dining, in sidewalk and street dining. However, it is Brooklyn that is far above the
# rest of the boroughs in rodent reports in 2020. This suggests that perhaps another factor is contributing
# to the rat problem in the Brooklyn borough. If restaurants were fully to blame for itβs rat problem, we
# would expect to see it having high numbers of restaurants approved for outdoor street and sidewalk
# dining, a number comparable to the borough of Manhattan.
#
# The first bar plot displays the number of rodent related 311 reports in the year 2020 by borough.
# Brooklyn leads the way with well over 10,000 rodent related calls in the year, while the next closest
# borough, Manhattan, only has about 8,000 rodent related calls in the year. In the bar plots related to
# restaurants, we see Manhattan leads the way across the board. In the restaurants with outdoor dining,
# sidewalk and street dining, Manhattan has twice as many restaurants than any other borough. Because
# of this vast difference in the number of restaurants in Manhattan compared to the other boroughs, we
# will narrow our focus to Manhattan in the next visualization."),),
fluidRow(align = "center",
br(),
br(),
br(),
# code for generating these plots
# title: "brendan2"
# author: "Brendan Mapes"
# date: "4/17/2021"
# output: html_document
# ---
#
# ```{r setup, include=TRUE}
# knitr::opts_chunk$set(echo = TRUE)
# library(ggplot2)
# library(ggthemes)
# library(gridExtra)
# library(dplyr)
# library(plotly)
# open <- read.csv("Open_Restaurant_Applications.csv")
# inspection <- read.csv("DOHMH_New_York_City_Restaurant_Inspection_Results.csv")
# rat_311 <- read.csv("311_Service_Requests_from_2010_to_Present.csv")
# restaurant_borough <- count(open, Borough)
# names(restaurant_borough)[names(restaurant_borough) == "n"] <- "count"
# rat_borough <- count(rat_311, Borough)
# borough <- c("Bronx", "Brooklyn", "Manhattan", "Queens", "Staten Island", "none")
# rat_borough <- cbind(rat_borough, borough) %>% filter(borough!= "none")
# rat_borough <- select(rat_borough, borough, n)
# names(rat_borough)[names(rat_borough) == "n"] <- "count"
# names(rat_borough)[names(rat_borough) == "borough"] <- "Borough"
# inspection_bc <- inspection %>% filter(GRADE == "B" | GRADE == "C")
# inspection_count_2020 <- count(inspection_bc, BORO)
# names(inspection_count_2020)[names(inspection_count_2020) == "n"] <- "count"
# names(inspection_count_2020)[names(inspection_count_2020) == "BORO"] <- "Borough"
# street_seating <- filter(open, Approved.for.Roadway.Seating == "yes")
# count_street <- count(street_seating, Borough)
# names(count_street)[names(count_street) == "n"] <- "count"
# sidewalk_seating <- filter(open, Approved.for.Sidewalk.Seating == "yes")
# count_sidewalk <- count(sidewalk_seating, Borough)
# names(count_sidewalk)[names(count_sidewalk) == "n"] <- "count"
# plot1 <- ggplot() + geom_bar(data=rat_borough, aes(x=Borough, y=count, fill=Borough), stat="identity") + ggtitle("2020 rodent reports") + ylab("Number of 311 calls\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot2 <- ggplot() + geom_bar(data=restaurant_borough, aes(x =Borough, y= count, fill =Borough), stat="identity") + ggtitle("Outdoor restaurants") + ylab("Applications approved\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), legend.position="none", axis.title.y=element_text(face="bold")) + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot3 <- ggplot() + geom_bar(data=inspection_count_2020, aes(x=Borough, y=count, fill=Borough), stat="identity") + ggtitle("Restaurants w/ B or C inspection scores") + ylab("Restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), legend.position="none", axis.title.y=element_text(face="bold")) + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot4 <- ggplot() + geom_bar(data=count_street, aes(x=Borough, y=count, fill=Borough), stat="identity") + ggtitle("Street dining") + ylab("Approved restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot5 <- ggplot() + geom_bar(data=count_sidewalk, aes(x=Borough, y=count, fill=Borough), stat="identity") + ggtitle("Sidewalk dining") + ylab("Approved restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
# plot1a <- ggplotly(plot1, tooltip=c("x", "y"))
# plot2a <- ggplotly(plot2, tooltip=c("x", "y"))
# plot3a <- ggplotly(plot3, tooltip=c("x", "y"))
# plot4a <- ggplotly(plot4, tooltip=c("x", "y"))
# plot5a <- ggplotly(plot5, tooltip=c("x", "y"))
# plot1a
# plot2a
# plot3a
# plot4a
# plot5a
h6("Chart 9"),
plotlyOutput("brendan_chart1" ,width="50%"),
br(),
h6("Chart 10"),
plotlyOutput("brendan_chart2",width="50%"),
br(),
h6("Chart 11"),
plotlyOutput("brendan_chart3",width='50%'),
br(),
h6("Chart 12"),
plotlyOutput("brendan_chart4",width='50%'),
br(),
h6("Chart 13"),
plotlyOutput("brendan_chart5",width='50%'),
br(),
p(class = "padding", align = "left", "In all five figures, we can see that Manhattan is far above the rest of the boroughs
in restaurants approved for outdoor dining, in sidewalk and street dining, and B and C graded restaurants. However, it is
Brooklyn that is far above the rest of the boroughs in rodent reports in 2020. This suggests that perhaps another factor is
contributing to the rat problem in the Brooklyn borough. If restaurants were fully to blame for itβs rat problem, we would
expect to see it having high numbers of restaurants approved for outdoor street and sidewalk dining, a number comparable to
the borough of Manhattan. The first bar plot displays the number of rodent related 311 reports in the year 2020 by borough.
Brooklyn leads the way with well over 10,000 rodent related calls in the year, while the next closest borough, Manhattan, only
has about 8,000 rodent related calls in the year. In the bar plots related to restaurants, we see Manhattan leads the way across
the board. In the restaurants with outdoor dining, sidewalk and street dining, Manhattan has twice as many restaurants than any
other borough. Because of this vast difference in the number of restaurants in Manhattan compared to the other boroughs, we will
narrow our focus to Manhattan in the next visualization."),br(),br(),
),
),
br(),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings", href="https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9"), h6("Filtered to 311 calls from Jan 1 2020 to Dec 31 2020, of complaint type βrodentβ in Manhattan borough.")),
h5(a("Open Restaurants", href="https://data.cityofnewyork.us/Transportation/Open-Restaurant-Applications/pitm-atqc/data"),h6("Filtered to Manhattan borough.")),
),
div(class='padding',
h4("Visualization:"),
),
fluidRow(align = "center",
h3("Rat sightings in 2020 overlaid on restaurant locations in Manhattan"),
h6("Chart 14"),),br(),
fluidRow(
leafletOutput("brendan_map", height = 500, width = "100%"),
br(),
br(),
p(class = "padding", align = "left", "Based off exploratory analysis of the restaurants and rat reports across all boroughs, itβs clear
Manhattanβs restaurant industry may be most closely linked to the rat problem than in other boroughs.
For that reason, we have provided an interactive map visualization of the Manhattan borough
specifically. In the map, restaurants are plotted and viewers can see the location and name of the
restaurant, along with whether or not the restaurant is available for open street or sidewalk dining. Also
charted on the map are the location of rat sightings in the 2020 311 calls data set, the same data used
for previous visualizations. With no zoom, clusters of the rats are displayed in the visualization. After
zooming in further, those clusters break into the individual rat sighting locations of which they consist.
Rat sighting locations are represented by small rat icons on the map."),
p(class = "padding", align = "left", "This visualization allows viewers to identify densely rat populated locations on the map and relate that
information to the data provided for restaurants in the same locations. In real time, such a map would
be useful for avoidance of rat βhot spotsβ when choosing places to dine. It also allows users to explore
which restaurants have practices that may be contributing to the rat problem the most, with lots of rat
sightings nearby."),),
fluidRow(
tags$style(".padding {
margin-left:30px;
margin-right:30px;
}"),
tags$style(".leftAlign{float:left;}"),
align = "left",
div(class='padding',
h4("Data used:"),
h5(a("Rat sightings", href="https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9"),
h6("Filtered to 311 calls from Jan 1 2020 to Dec 31 2020, of complaint type 'rodent' in Manhattan borough.
Calls related to noise, parking violations, and other non-emergency police related matters are also
excluded from this visualization.")),
),
div(class='padding',
h4("Visualization:"),
),
),
fluidRow(align = "center",
h3("Wordcloud of the descriptor variable of all NYC311 complaints in 2020"),
br(),
h6("Chart 15"),
img(src='Capture.PNG',width="50%"),
br(),
h6("Chart 16"),
img(src='Picture2.png',width="50%"),
br(),
h6("For reference:"),
img(src='rat.png',),
br(),
br(),
p(class = "padding", align = "left", "For required text analysis, we have again referred to the 2020 rodent related 311 reports, specifically on
the descriptor variable, where various notes are left on the nature or details of the complaint. Two word
clouds are presented. The first is a basic wordcloud created with the ggwordcloud package. Words
related to rat sightings are differentiated from others by color. Viewers can see in this wordcloud that
the descriptor variable does have lots of mentions of rodent related issues. The second wordcloud
presented is also from the ggwordcloud package, but with an added mask, intended to create a
wordcloud the shape of a rat. This visualization is slightly more visually appealing, but reveals the exact
same information to the reader. Rat sightings are often mentioned in the descriptor variable of the data
set."),br(),br()
),
# fluidRow(
# align = "center",
# plotOutput("brendan_wc1"),
# plotOutput("bendan_wc2"),
# ),
),
)
)
# code for generating the plots -----------------------------------------------------------------
server <- function(input, output, session) {
# prajwal's code for generating interactive map backend -------------------
# points <- eventReactive(input$recalc, {
# cbind(rat_sightings_sample$latitude, rat_sightings_sample$longitude)
# }, ignoreNULL = FALSE)
#
observe({
min_year <- input$year_input[1]
max_year <- input$year_input[2]
burough <- input$burough_input
case_status1 <- input$case_status
rat_sightings_sample <- rat_sightings[sample.int(nrow(rat_sightings), input$num_sample),]
#rat_sightings_sample <- rat_sightings
latitude_colnum <- grep('latitude', colnames(rat_sightings_sample))
longitude_colnum <- grep('longitude', colnames(rat_sightings_sample))
rat_sightings_sample <- rat_sightings_sample[complete.cases(rat_sightings_sample[,latitude_colnum:longitude_colnum]),]
rat_sightings_sample$year_created <- year(parse_date_time(rat_sightings_sample$Created.Date, '%m/%d/%y %I:%M:%S %p'))
#print('buroughh')
#print(burough)
#filter_rat_sightings <- rat_sightings_sample %>% filter(year_created >= min_year, year_created <= max_year, Borough %in% burough)
check_rows_of_filter <- nrow(rat_sightings_sample %>% filter(year_created >= min_year, year_created <= max_year,
Borough %in% burough,
Status %in% case_status1))
rat_sightings_buroughs2 <- as.character(unique(unlist(rat_sightings_sample$Borough)))
rat_sightings_case_status2 <- as.character(unique(unlist(rat_sightings_sample$Status)))
# print('buroughs 2')
# print(rat_sightings_buroughs2)
# print('case statuses 2')
# print(rat_sightings_case_status2)
if (check_rows_of_filter <= 0){
#updateSliderInput(session, "year_input", value = c(2010, 2021))
#updateSliderInput(session, "burough_input", value = rat_sightings_buroughs2)
#updateSliderInput(session, "case_status", value = rat_sightings_case_status2)
# filter_rat_sightings2 <- rat_sightings_sample
# reset("year_input")
# reset("burough_input")
# reset("burough_input")
# print('in the case of 0 rows, resetting to entire df')
leaflet() %>%
addProviderTiles(providers$Stamen.TonerLite,
options = providerTileOptions(noWrap = TRUE)
) %>%
setView(lng = -73.98928, lat = 40.75042, zoom = 10)
}
else{
filter_rat_sightings2 <- rat_sightings_sample %>% filter(year_created >= min_year, year_created <= max_year,
Borough %in% burough,
Status %in% case_status1)
filter_rat_sightings <- filter_rat_sightings2
getColor <- function(filter_rat_sightings, i) {
if(filter_rat_sightings$Status[i] == "Closed") {
"green"
}
else if(filter_rat_sightings$Status[i] == "In Progress") {
"lightblue"
}
else if(filter_rat_sightings$Status[i] == "Assigned") {
"orange"
}
else if(filter_rat_sightings$Status[i] == "Open") {
"purple"
}
else if(filter_rat_sightings$Status[i] == "Pending") {
"darkred"
}
else if(filter_rat_sightings$Status[i] == "Draft") {
"blue"
}}
markerColors <- rep(NA, nrow(filter_rat_sightings))
for (i in 1:nrow(filter_rat_sightings)){
markerColors[i] <- getColor(filter_rat_sightings, i)
}
icons <- awesomeIcons(
icon = 'ios-close',
iconColor = 'cadetblue',
library = 'ion',
markerColor = markerColors
)
output$map <- renderLeaflet({
leaflet(data = filter_rat_sightings) %>%
addProviderTiles(providers$Stamen.TonerLite,
options = providerTileOptions(noWrap = TRUE)
) %>%
setView(lng = -73.98928, lat = 40.75042, zoom = 10) %>%
addAwesomeMarkers( ~longitude, ~latitude, clusterOptions = markerClusterOptions() ,icon = icons,
popup = as.character(paste('Created date:', filter_rat_sightings$Created.Date,'<br>',
'Closed Date:', filter_rat_sightings$Closed.Date,'<br>',
#'Complaint type:',filter_rat_sightings$Complaint.Type,'<br>',
#'Descriptor:',filter_rat_sightings$Descriptor,'<br>',
'Address:',filter_rat_sightings$Incident.Address,'<br>',
'Status:', filter_rat_sightings$Status, '<br>',
'Location Type:', filter_rat_sightings$Location.Type))) %>%
addHeatmap( ~longitude, ~latitude, group = "heat",max=1, blur = 45, minOpacity = 0.8) %>% addLegend("topleft",
colors =c('green', "darkred", "lightblue", "orange"),
#"purple", "#56a0d1"
labels= c("Closed", "Pending", "In Progress","Assigned"),
# "Open", "Draft"
title= "Complaint status",
opacity = 1)
})
output$cityViz <- renderPlotly({
if (nrow(zipsInBounds()) == 0)
return(NULL)
tmp <- (zipsInBounds() %>% count(City))
tmp <- tmp[order(-tmp$n),]
tmp <- tmp[1:5,]
ggplotly(
ggplot(tmp, aes(x=City, y=n, fill = City)) + geom_bar(stat="identity") + ylab("Top 5 visible buroughs") + theme(legend.position = "none") + scale_color_brewer(palette="Dark2")+
theme(axis.title.x=element_blank(),
axis.ticks.x=element_blank())
)
})
output$locationViz <- renderPlotly({
if (nrow(zipsInBounds()) == 0)
return(NULL)
tmp <- (zipsInBounds() %>% count(Location.Type))
tmp <- tmp[order(-tmp$n),]
tmp <- tmp[1:5,]
ggplotly(tooltip = c("n"),
ggplot(tmp, aes(x=Location.Type, y=n)) + geom_bar(stat="identity", aes(fill = Location.Type)) + ggtitle('Visible location types') + ylab("Visible location types") +
theme_light()+
theme(axis.title.y=element_blank(),
#axis.text.y=element_blank(),
axis.title.x=element_blank(),
axis.text.x=element_blank(),
axis.ticks.x=element_blank()) + labs(fill = "Visible location types") + coord_flip() + theme(legend.position = "none")
)
})
output$yearViz <- renderPlotly({
if (nrow(zipsInBounds()) == 0)
return(NULL)
# total cases
created_date_sample <- data.table(zipsInBounds()$Created.Date)
created_date_sample$dates <- parse_date_time(created_date_sample$V1, '%m/%d/%y %I:%M:%S %p')
plot_created_year <- data.frame(table(year(date(created_date_sample$dates))))
for (i in 2010:2021){
if ((i %in% plot_created_year$Var1)==FALSE) {
#print(i)
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- c('Var1','Freq')
plot_created_year <- rbind(plot_created_year, tmp_df)
}
}
plot_created_year$Var1 <- as.numeric(as.character(plot_created_year$Var1))
names(plot_created_year)[names(plot_created_year) == "Var1"] <- "Year"
plot_created_year <- plot_created_year[order(plot_created_year$Year),]
plot_created_year <- filter(plot_created_year, Year >= min_year, Year <= max_year)
plot_created_year$case_status <- 'Total'
# closed cases
plot_closed<- filter(zipsInBounds(), Status == 'Closed')
flag_closed <- 0
if (nrow(plot_closed) == 0) {
flag_closed <- 1
}
if (flag_closed == 0){
plot_closed <- data.table(plot_closed$Created.Date)
plot_closed$dates <- parse_date_time(plot_closed$V1, '%m/%d/%y %I:%M:%S %p')
plot_closed_year <- data.frame(table(year(date(plot_closed$dates))))
#print(plot_closed_year)
for (i in 2010:2021){
if ((i %in% plot_closed_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_closed_year)
plot_closed_year <- rbind(plot_closed_year, tmp_df)
}
}
plot_closed_year$Var1 <- as.numeric(as.character(plot_closed_year$Var1))
names(plot_closed_year)[names(plot_closed_year) == "Var1"] <- "Year"
plot_closed_year <- plot_closed_year[order(plot_closed_year$Year),]
plot_closed_year <- filter(plot_closed_year, Year >= min_year, Year <= max_year)
plot_closed_year$case_status <- 'Closed'
}
# assigned cases
plot_assigned<- filter(zipsInBounds(), Status == 'Assigned')
flag_assigned <- 0
if (nrow(plot_assigned) == 0) {
flag_assigned <- 1
}
if (flag_assigned == 0){
plot_assigned <- data.table(plot_assigned$Created.Date)
plot_assigned$dates <- parse_date_time(plot_assigned$V1, '%m/%d/%y %I:%M:%S %p')
plot_assigned_year <- data.frame(table(year(date(plot_assigned$dates))))
#print(plot_assigned_year)
for (i in 2010:2021){
if ((i %in% plot_assigned_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_assigned_year)
plot_assigned_year <- rbind(plot_assigned_year, tmp_df)
}
}
plot_assigned_year$Var1 <- as.numeric(as.character(plot_assigned_year$Var1))
names(plot_assigned_year)[names(plot_assigned_year) == "Var1"] <- "Year"
plot_assigned_year <- plot_assigned_year[order(plot_assigned_year$Year),]
plot_assigned_year <- filter(plot_assigned_year, Year >= min_year, Year <= max_year)
plot_assigned_year$case_status <- 'Assigned'
#print('assigned or in progress')
#print(plot_assigned_year)
}
# assigned or in progress cases
plot_in_progress<- filter(zipsInBounds(), Status == 'In Progress')
flag_in_progress <- 0
if (nrow(plot_in_progress) == 0) {
flag_in_progress <- 1
}
if (flag_in_progress == 0){
plot_in_progress <- data.table(plot_in_progress$Created.Date)
plot_in_progress$dates <- parse_date_time(plot_in_progress$V1, '%m/%d/%y %I:%M:%S %p')
plot_in_progress_year <- data.frame(table(year(date(plot_in_progress$dates))))
#print(plot_in_progress_year)
for (i in 2010:2021){
if ((i %in% plot_in_progress_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_in_progress_year)
plot_in_progress_year <- rbind(plot_in_progress_year, tmp_df)
}
}
plot_in_progress_year$Var1 <- as.numeric(as.character(plot_in_progress_year$Var1))
names(plot_in_progress_year)[names(plot_in_progress_year) == "Var1"] <- "Year"
plot_in_progress_year <- plot_in_progress_year[order(plot_in_progress_year$Year),]
plot_in_progress_year <- filter(plot_in_progress_year, Year >= min_year, Year <= max_year)
plot_in_progress_year$case_status <- 'In Progress'
#print('assigned or in progress')
#print(plot_in_progress_year)
}
# open cases
plot_open<- filter(zipsInBounds(), Status == 'Open')
flag_open <- 0
if (nrow(plot_open) == 0) {
flag_open <- 1
}
if (flag_open == 0){
plot_open <- data.table(plot_open$Created.Date)
plot_open$dates <- parse_date_time(plot_open$V1, '%m/%d/%y %I:%M:%S %p')
plot_open_year <- data.frame(table(year(date(plot_open$dates))))
#print(plot_open_year)
for (i in 2010:2021){
if ((i %in% plot_open_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_open_year)
plot_open_year <- rbind(plot_open_year, tmp_df)
}
}
plot_open_year$Var1 <- as.numeric(as.character(plot_open_year$Var1))
names(plot_open_year)[names(plot_open_year) == "Var1"] <- "Year"
plot_open_year <- plot_open_year[order(plot_open_year$Year),]
plot_open_year <- filter(plot_open_year, Year >= min_year, Year <= max_year)
plot_open_year$case_status <- 'Open'
#print('open or pending')
#print(plot_open_year)
# print('created')
# print(plot_created_year)
# print('closed')
# print(plot_closed_year)
# print('combined')
}
# pending cases
plot_pending<- filter(zipsInBounds(), Status == 'Pending')
flag_pending <- 0
if (nrow(plot_pending) == 0) {
flag_pending <- 1
}
if (flag_pending == 0){
plot_pending <- data.table(plot_pending$Created.Date)
plot_pending$dates <- parse_date_time(plot_pending$V1, '%m/%d/%y %I:%M:%S %p')
plot_pending_year <- data.frame(table(year(date(plot_pending$dates))))
#print(plot_pending_year)
for (i in 2010:2021){
if ((i %in% plot_pending_year$Var1)==FALSE) {
tmp_df <- data.frame(toString(i), 0)
names(tmp_df) <- colnames(plot_pending_year)
plot_pending_year <- rbind(plot_pending_year, tmp_df)
}
}
plot_pending_year$Var1 <- as.numeric(as.character(plot_pending_year$Var1))
names(plot_pending_year)[names(plot_pending_year) == "Var1"] <- "Year"
plot_pending_year <- plot_pending_year[order(plot_pending_year$Year),]
plot_pending_year <- filter(plot_pending_year, Year >= min_year, Year <= max_year)
plot_pending_year$case_status <- 'Pending'
#print('open or pending')
#print(plot_pending_year)
# print('created')
# print(plot_created_year)
# print('closed')
# print(plot_closed_year)
# print('combined')
}
#plot_this <- plot_created_year
plot_this <- data.frame(matrix(ncol = 3, nrow = 0))
colnames(plot_this) <- c("Year",'Freq','case_status')
rownames(plot_created_year)
# print('flag open')
# print(flag_open)
# print('flag pending')
# print(flag_pending)
# print('flag assigned')
# print(flag_assigned)
# print('flag in progress')
# print(flag_in_progress)
# print('flag closed')
# print(flag_closed)
if (flag_open == 0){
plot_this <- rbind(plot_this, plot_open_year)
}
if (flag_pending == 0){
plot_this <- rbind(plot_this, plot_pending_year)
}
if (flag_assigned == 0) {
plot_this <- rbind(plot_this, plot_assigned_year)
}
if (flag_in_progress == 0) {
plot_this <- rbind(plot_this, plot_in_progress_year)
}
if (flag_closed == 0){
plot_this <- rbind(plot_this, plot_closed_year)
}
#print(plot_this)
# p_years <- ggplotly(
# ggplot(data=plot_created_year, aes(x=Year, y=Freq)) + geom_path(stat="identity") + ylab('Rat sightings') + geom_point()+
# theme(axis.title.x=element_blank()) + scale_x_continuous(breaks=seq(min_year, max_year, 1))
# )
colors_for_case_status <- rep(NA, nrow(plot_this))
for (i in 1:nrow(plot_this)){
if (plot_this$case_status[i] == "Closed"){
colors_for_case_status[i] <- "green"
}
else if (plot_this$case_status[i] == "Pending"){
colors_for_case_status[i] <- "darkred"
}
else if (plot_this$case_status[i] == "Assigned"){
colors_for_case_status[i] <- "orange"
}
else if (plot_this$case_status[i] == "In Progress"){
colors_for_case_status[i] <- "cadetblue"
}
else if (plot_this$case_status[i] == "Draft"){
colors_for_case_status[i] <- "blue"
}
else if (plot_this$case_status[i] == "Open"){
colors_for_case_status[i] <- "purple"
}
}
p_years <- ggplotly(
ggplot(data=plot_this, aes(x=Year, y=Freq)) + geom_line(aes(color=case_status)) + geom_point(aes(colour=case_status))
#+ scale_colour_manual(name = 'Case status',values =c('green'='green','cadetblue' = 'cadetblue', 'orange'='orange', 'darkred'='darkred'), labels = c("closed","in progress", "assigned",'pending'))
+ ggtitle('Complaint status trend') + scale_x_continuous(breaks=seq(min_year, max_year, 1)) + theme_light()
+ theme(axis.title.y=element_blank(),
#axis.text.y=element_blank(),
axis.title.x=element_blank(),)
#axis.text.x=element_blank())
+ theme(legend.title = element_blank())
#+ theme(legend.position="left")
#+ labs(color='Status')
+
theme(axis.title.x=element_blank())
)
})
zipsInBounds <- reactive({
if (is.null(input$map_bounds))
return(zipdata[FALSE,])
bounds <- input$map_bounds
#print(bounds)
latRng <- range(bounds$north, bounds$south)
lngRng <- range(bounds$east, bounds$west)
#print(latRng)
subset(filter_rat_sightings,
latitude >= latRng[1] & latitude <= latRng[2] &
longitude >= lngRng[1] & longitude <= lngRng[2])
})
}
#filter_rat_sightings <- filter_rat_sightings[,burough]
# if (nrow(event_data("plotly_selecting"))>0){
# filter_rat_sightings <- filter_rat_sightings %>% filter(year_created %in% event_data("plotly_selecting")$Var1)
# }
#print('reached here')
#print(filter_rat_sightings)
#print('end reached here')
})
# PRATISHTA'S VISUALIZATIONS -------------------------------------------------
output$pratishta1 <- renderPlotly({
p <- ggplot(rat_ton_date, aes(x=Created.Date, y=REFUSETONSCOLLECTED)) +
geom_line(aes(color = Borough)) +
geom_point(aes(color = Borough)) +
xlab("Date by Months") +
ylab("Weight of Waste (Tons)") + theme_light()
p
})
output$pratishta2 <- renderPlotly({
p <- ggplot(rat_ton_date, aes(x=Created.Date, y=n)) +
geom_line(aes(color = Borough)) +
geom_point(aes(color = Borough)) +
xlab("Date by Months") +
ylab("Number of rat sightings") + theme_light()
p
})
output$pratishta3 <- renderPlotly({
p <- ggplot(rat_ton_date, aes(x=Created.Date, y=rate)) +
geom_line(aes(color = Borough)) +
geom_point(aes(color = Borough)) +
xlab("Date by Months") +
ylab("Rate of rats per kiloton of waste") + theme_light()
p
})
output$pratishta4 <- renderPlotly({
ton +
xlab("Boroughs") +
ylab("Weight of Waste (Tons)") + theme_light()
})
output$pratishta5 <- renderPlotly({
rat +
xlab("Boroughs") +
ylab("Number of Rat Sightings") + theme_light()
})
output$pratishta7 <- renderTmap({
## ------------------------------------------------------------------------
tm_shape(nyc_sp) +
tm_fill("n", title = "Rat Sightings in Community Districts")+tm_view(set.view = c(-73.98928, 40.70042,10))
})
output$pratishta8 <- renderTmap({
## ------------------------------------------------------------------------
tm_shape(nyc_sp) +
tm_fill("total_ton", title = "Tones of Waste and Rat Sightings by DSNY Districts")+tm_view(set.view = c(-73.98928, 40.70042,10))
})
output$pratishta9 <- renderPlot({
## ------------------------------------------------------------------------
legend_creator = function(col.regions, xlab, ylab, nbins){
bilegend = levelplot(matrix(1:(nbins * nbins), nrow = nbins),
axes = FALSE, col.regions = col.regions,
xlab = xlab, ylab = ylab,
cuts = 8, colorkey = FALSE, scales = list(draw = 0))
bilegend
}
add_new_var = function(x, var1, var2, nbins, style = "quantile"){
class1 = suppressWarnings(findCols(classIntervals(c(x[[var1]]),
n = nbins,
style = style)))
class2 = suppressWarnings(findCols(classIntervals(c(x[[var2]]),
n = nbins,
style = style)))
x$new_class = class1 + nbins * (class2 - 1)
return(x)
}
## ------------------------------------------------------------------------
nyc_cd <- nyc_sp
## ------------------------------------------------------------------------
nyc_cd = add_new_var(nyc_cd,
var1 = "n",
var2 = "total_ton",
nbins = 3)
## ------------------------------------------------------------------------
bilegend = legend_creator(stevens.pinkblue(n = 9),
xlab = "rats",
ylab = "total tonnes",
nbins = 3)
vp = viewport(x = 0.25, y = 0.25, width = 0.25, height = 0.25)
pushViewport(vp)
#print(bilegend, newpage = FALSE)
bimap = tm_shape(nyc_cd) +
tm_fill("new_class", style = "cat", palette = stevens.pinkblue(n = 9), legend.show = FALSE) +
tm_layout(legend.show = FALSE)
grid.newpage()
vp = viewport(x = 0.37, y = 0.75, width = 0.25, height = 0.25)
print(bimap, vp = viewport())
pushViewport(vp)
print(bilegend, newpage = FALSE, vp = vp)
})
# END OF PRATISHTA'S VISUALIZATIONS ------------------------------------------
# brendan's viz -----------------------------------------------------------
output$brendan_chart1 <- renderPlotly({
plot1 <-ggplot() + geom_bar(data=rat_borough, aes(x=Borough, y=count), stat="identity", fill=rainbow(n=5)) + ggtitle("2020 rodent reports") + ylab("Number of 311 calls\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot1
})
output$brendan_chart2 <- renderPlotly({
plot2 <- ggplot() + geom_bar(data=restaurant_borough, aes(x =Borough, y= count), stat="identity", fill=rainbow(n=5)) + ggtitle("Outdoor restaurants") + ylab("Applications approved\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), legend.position="none", axis.title.y=element_text(face="bold")) + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot2
})
output$brendan_chart3 <- renderPlotly({
plot3 <- ggplot() + geom_bar(data=inspection_count_2020, aes(x=Borough, y=count), stat="identity", fill=rainbow(n=5)) + ggtitle("Restaurants w/ B or C inspection scores") + ylab("Restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), legend.position="none", axis.title.y=element_text(face="bold")) + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot3
})
output$brendan_chart4 <- renderPlotly({
plot4 <- ggplot() + geom_bar(data=count_street, aes(x=Borough, y=count), stat="identity", fill=rainbow(n=5)) + ggtitle("Street dining") + ylab("Approved restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot4
})
output$brendan_chart5 <- renderPlotly({
plot5 <- ggplot() + geom_bar(data=count_sidewalk, aes(x=Borough, y=count), stat="identity", fill=rainbow(n=5)) + ggtitle("Sidewalk dining") + ylab("Approved restaurants\n") + xlab("\nBorough") + theme_light() + theme(plot.title=element_text(face="bold"), axis.title.x=element_text(face="bold"), axis.title.y=element_text(face="bold"), legend.position="none") + theme(axis.text.x = element_text(angle = 40, vjust=.7))
plot5
})
output$brendan_map <- renderLeaflet({
interactive_map <- leaflet() %>% addProviderTiles(providers$Stamen.TonerLite,
options = providerTileOptions(noWrap = TRUE)) %>% addCircleMarkers(data=manhattan_open, lng=~Longitude, lat=~Latitude, radius=1, color= "red", fillOpacity=1, popup=~paste('Restuarant:', manhattan_open$Restaurant.Name, '<br>', 'Street Dining?', manhattan_open$Approved.for.Roadway.Seating, '<br>', 'Sidewalk Dining?', manhattan_open$Approved.for.Sidewalk.Seating, '<br>')) %>% addMarkers(data=manhattan_311,lng=~manhattan_311.Longitude, lat=~manhattan_311.Latitude, icon=icon, popup=~paste('Address:',manhattan_311$manhattan_311.Incident.Address,'<br>', 'Call Date:', manhattan_311$manhattan_311.Created.Date,'<br>','Descriptor:', manhattan_311$manhattan_311.Descriptor,'<br>'), clusterOptions= markerClusterOptions(disableClusteringAtZoom=16)) %>%
setView(lng = -73.98928, lat = 40.77042, zoom = 12)
})
# output$brendan_wc1 <- renderPlot({
# (wordcloud1 <- ggplot(descriptors3, aes(label=word, size=n, color=col)) + geom_text_wordcloud_area(mask=img, rm_outside = TRUE) + scale_size_area(max_size=5) + theme_classic())
#
# })
#
# output$brendan_wc2 <- renderPlot({
# (wordcloud2 <- ggplot(descriptors3, aes(label=word, size=n, color=col)) + geom_text_wordcloud_area() + scale_size_area())
#
# })
}
shinyApp(ui = ui, server = server)
|
home <- tabPanel("Home",
sidebarLayout(sidebarPanel(
p(
"Hi there, my name is Bayan Hammami
and I'm a data scientist who loves data.
This is my interactive data-driven resume.
I hope you enjoy it!"
),
p(
"This site was built with R and the following packages;
shiny, ggplot2, ggnet, network, wordcloud, fmsb, dplyr, tm and a few more."
),
p(
"The code for this app can be found here:", tags$b(tags$a(href ="https://github.com/BayanHammami/resume-shiny", "https://github.com/BayanHammami/resume-shiny"))
),
p(
"Please contact me on:", tags$b("bayan.hammami@gmail.com")
),
hr(),
sliderInput(
"min_word_freq",
"Minimum word frequency in my text resume",
min = 1,
max = 5,
value = 2
),
actionButton("home_regenerate", "Regenerate Wordcloud"),
hr(),
span(
style="margin:8px;",
a(href="https://github.com/BayanHammami",
target="_blank",
icon("github", class = "fa-2x", lib = "font-awesome")
)
),
span(
style="margin:8px;",
a(href="https://www.linkedin.com/in/bayan-hammami",
target="_blank",
icon("linkedin", class = "fa-2x", lib = "font-awesome")
)
),
span(
style="margin:8px;",
a(href="https://twitter.com/bayan_aus",
target="_blank",
icon("twitter", class = "fa-2x", lib = "font-awesome")
)
)
),
mainPanel(
plotOutput("home_plot", height = "600px"),
style="text-align: center;"
))
)
| /ui-components/nav-panels/ui-home.R | no_license | BayanHammami/resume-shiny | R | false | false | 2,557 | r | home <- tabPanel("Home",
sidebarLayout(sidebarPanel(
p(
"Hi there, my name is Bayan Hammami
and I'm a data scientist who loves data.
This is my interactive data-driven resume.
I hope you enjoy it!"
),
p(
"This site was built with R and the following packages;
shiny, ggplot2, ggnet, network, wordcloud, fmsb, dplyr, tm and a few more."
),
p(
"The code for this app can be found here:", tags$b(tags$a(href ="https://github.com/BayanHammami/resume-shiny", "https://github.com/BayanHammami/resume-shiny"))
),
p(
"Please contact me on:", tags$b("bayan.hammami@gmail.com")
),
hr(),
sliderInput(
"min_word_freq",
"Minimum word frequency in my text resume",
min = 1,
max = 5,
value = 2
),
actionButton("home_regenerate", "Regenerate Wordcloud"),
hr(),
span(
style="margin:8px;",
a(href="https://github.com/BayanHammami",
target="_blank",
icon("github", class = "fa-2x", lib = "font-awesome")
)
),
span(
style="margin:8px;",
a(href="https://www.linkedin.com/in/bayan-hammami",
target="_blank",
icon("linkedin", class = "fa-2x", lib = "font-awesome")
)
),
span(
style="margin:8px;",
a(href="https://twitter.com/bayan_aus",
target="_blank",
icon("twitter", class = "fa-2x", lib = "font-awesome")
)
)
),
mainPanel(
plotOutput("home_plot", height = "600px"),
style="text-align: center;"
))
)
|
###tidy_tuesday
###4/10/21
###richard rachman
#######################
### library
library(tidyverse)
library(tidytuesdayR)
library(here)
library(viridis)
library(hrbrthemes)
### data
forest <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-04-06/forest.csv')
view(forest)
entity<- forest$entity
NorthAmerica<- c("Canada","Cuba","Mexico","United States")
forest1<- filter(forest, entity %in% NorthAmerica) %>%
select(entity, year, net_forest_conversion)
# North American countries
# Belize
# Canada
# Costa Rica
# Cuba
# Dominican Republic
# El Salvador
# Guatemala
# Honduras
# Jamaica
# Mexico
# Nicaragua
# Panama
# USA
### script
ggplot(forest1, aes(year, net_forest_conversion, label= entity, color= entity)) +
geom_line((aes(colour = entity)))+
scale_color_viridis(discrete = TRUE) +
ggtitle("Forest conversion by largest countries in NA") +
theme_ipsum() +
ylab("Scale of net forest conversion")+
xlab("Year")+
labs(caption = "Positive forest coversion suggestions forest expansion,
though data is incomplete for USA")+
theme(legend.title = element_blank())+
ggsave(here("output","forests.jpeg"))
# Plot
| /script/tidytuesday_4-12-21.R | no_license | awanderingecologist/tidytuesday | R | false | false | 1,211 | r | ###tidy_tuesday
###4/10/21
###richard rachman
#######################
### library
library(tidyverse)
library(tidytuesdayR)
library(here)
library(viridis)
library(hrbrthemes)
### data
forest <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-04-06/forest.csv')
view(forest)
entity<- forest$entity
NorthAmerica<- c("Canada","Cuba","Mexico","United States")
forest1<- filter(forest, entity %in% NorthAmerica) %>%
select(entity, year, net_forest_conversion)
# North American countries
# Belize
# Canada
# Costa Rica
# Cuba
# Dominican Republic
# El Salvador
# Guatemala
# Honduras
# Jamaica
# Mexico
# Nicaragua
# Panama
# USA
### script
ggplot(forest1, aes(year, net_forest_conversion, label= entity, color= entity)) +
geom_line((aes(colour = entity)))+
scale_color_viridis(discrete = TRUE) +
ggtitle("Forest conversion by largest countries in NA") +
theme_ipsum() +
ylab("Scale of net forest conversion")+
xlab("Year")+
labs(caption = "Positive forest coversion suggestions forest expansion,
though data is incomplete for USA")+
theme(legend.title = element_blank())+
ggsave(here("output","forests.jpeg"))
# Plot
|
library(bigMap)
### Name: bdm.pakde
### Title: Perplexity-adaptive kernel density estimation
### Aliases: bdm.pakde
### ** Examples
# --- load mapped dataset
bdm.example()
# --- run paKDE
## Not run:
##D exMap <- bdm.pakde(exMap, threads = 4, ppx = 200, g = 200, g.exp = 3)
## End(Not run)
# --- plot paKDE output
bdm.plot(exMap)
| /data/genthat_extracted_code/bigMap/examples/bdm.pakde.Rd.R | no_license | surayaaramli/typeRrh | R | false | false | 339 | r | library(bigMap)
### Name: bdm.pakde
### Title: Perplexity-adaptive kernel density estimation
### Aliases: bdm.pakde
### ** Examples
# --- load mapped dataset
bdm.example()
# --- run paKDE
## Not run:
##D exMap <- bdm.pakde(exMap, threads = 4, ppx = 200, g = 200, g.exp = 3)
## End(Not run)
# --- plot paKDE output
bdm.plot(exMap)
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/clear_selection.R
\name{clear_selection}
\alias{clear_selection}
\title{Clear a selection of nodes or edges in a graph}
\usage{
clear_selection(graph)
}
\arguments{
\item{graph}{a graph object of class
\code{dgr_graph} that is created using
\code{create_graph}.}
}
\value{
a graph object of class \code{dgr_graph}.
}
\description{
Clear the selection of nodes or edges
within a graph object.
}
\examples{
# Create a simple graph
nodes <-
create_nodes(
nodes = c("a", "b", "c", "d"),
type = "letter",
label = TRUE,
value = c(3.5, 2.6, 9.4, 2.7))
edges <-
create_edges(
from = c("a", "b", "c"),
to = c("d", "c", "a"),
rel = "leading_to")
graph <-
create_graph(
nodes_df = nodes,
edges_df = edges)
# Select nodes "a" and "c"
graph <-
select_nodes(
graph = graph,
nodes = c("a", "c"))
# Verify that a node selection has been made
get_selection(graph)
#> [1] "a" "c"
# Clear the selection with `clear_selection()`
graph <- clear_selection(graph = graph)
# Verify that the node selection has been cleared
get_selection(graph)
#> [1] NA
}
| /man/clear_selection.Rd | no_license | timelyportfolio/DiagrammeR | R | false | true | 1,169 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/clear_selection.R
\name{clear_selection}
\alias{clear_selection}
\title{Clear a selection of nodes or edges in a graph}
\usage{
clear_selection(graph)
}
\arguments{
\item{graph}{a graph object of class
\code{dgr_graph} that is created using
\code{create_graph}.}
}
\value{
a graph object of class \code{dgr_graph}.
}
\description{
Clear the selection of nodes or edges
within a graph object.
}
\examples{
# Create a simple graph
nodes <-
create_nodes(
nodes = c("a", "b", "c", "d"),
type = "letter",
label = TRUE,
value = c(3.5, 2.6, 9.4, 2.7))
edges <-
create_edges(
from = c("a", "b", "c"),
to = c("d", "c", "a"),
rel = "leading_to")
graph <-
create_graph(
nodes_df = nodes,
edges_df = edges)
# Select nodes "a" and "c"
graph <-
select_nodes(
graph = graph,
nodes = c("a", "c"))
# Verify that a node selection has been made
get_selection(graph)
#> [1] "a" "c"
# Clear the selection with `clear_selection()`
graph <- clear_selection(graph = graph)
# Verify that the node selection has been cleared
get_selection(graph)
#> [1] NA
}
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/givenNamesDB_authorships.R
\docType{data}
\name{givenNamesDB_authorships}
\alias{givenNamesDB_authorships}
\title{Gender data for authorship sample}
\format{A data.table object with 872 rows and 4 variables:
\describe{
\item{name}{A term used as first name.}
\item{gender}{The predicted gender for the term.}
\item{probability}{The probability of the predicted gender.}
\item{count}{How many social profiles with the term as a given name
is recorded in the genderize.io database.}
}}
\source{
\url{http://genderize.io/}
}
\usage{
givenNamesDB_authorships
}
\description{
A dataset with first names and gender data from genderize.io for
\pkg{authorships} dataset in the package. This is the output
of \code{findGivenNames}
function that was performed on December 26, 2014.
}
\keyword{datasets}
| /man/givenNamesDB_authorships.Rd | permissive | cran/genderizeR | R | false | true | 913 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/givenNamesDB_authorships.R
\docType{data}
\name{givenNamesDB_authorships}
\alias{givenNamesDB_authorships}
\title{Gender data for authorship sample}
\format{A data.table object with 872 rows and 4 variables:
\describe{
\item{name}{A term used as first name.}
\item{gender}{The predicted gender for the term.}
\item{probability}{The probability of the predicted gender.}
\item{count}{How many social profiles with the term as a given name
is recorded in the genderize.io database.}
}}
\source{
\url{http://genderize.io/}
}
\usage{
givenNamesDB_authorships
}
\description{
A dataset with first names and gender data from genderize.io for
\pkg{authorships} dataset in the package. This is the output
of \code{findGivenNames}
function that was performed on December 26, 2014.
}
\keyword{datasets}
|
sidebar <- dashboardSidebar(
sidebarMenu(
menuItem("Webinar Participation", tabName = "social_location", icon = icon("user", lib = "glyphicon"),
badgeLabel = "main", badgeColor = "green"),
menuItem("User Journey", icon = icon("send", lib = "glyphicon"), tabName = "landing_page"
#,
#badgeLabel = "path", badgeColor = "orange"
)
# ,
# menuItem("valueBox test", tabName = "valueboxtest", icon = icon("thumbs-up", lib = "glyphicon"))
)
)
body <- dashboardBody(
tabItems(
tabItem(tabName = "social_location",
fluidRow(
column(
width = 5,
box(width = 12,
title = "DA webinar registrations - July 2020",
plotlyOutput("bulletgraph13", height = "120px")
)
),
column(width = 2,
br(),
br(),
valueBox(width = 12, goals$goal17Completions[5] + goals$goal13Completions[5], "Total signups (July)", icon = icon("list"))
),
column(width = 5,
box(width = 12,
title = "PSD webinar registrations - July 2020",
plotlyOutput("bulletgraph17", height = "120px")
)
)
),
fluidRow(
width = 12,
plotOutput("goals")
),
fluidRow(
column(width = 6,
br(),
plotOutput("social")
),
column(width = 6,
br(),
plotOutput("referrers")
)
),
),
tabItem(tabName = "landing_page",
fluidRow(
box(width = 12,
plotOutput("landing")
)
),
fluidRow(
box(width = 6,
plotOutput("landing_goals13")
),
box(width = 6,
plotOutput("landing_goals17")
)
)
)
# tabItem(tabName = "valueboxtest",
# valueBox(50*2, "valueBox test", icon = icon("list")))
)
)
ui <- dashboardPage(
dashboardHeader(title = "Website Engagement"),
sidebar,
body
)
| /ui.R | no_license | Debs20/data-proj-syn | R | false | false | 2,527 | r |
sidebar <- dashboardSidebar(
sidebarMenu(
menuItem("Webinar Participation", tabName = "social_location", icon = icon("user", lib = "glyphicon"),
badgeLabel = "main", badgeColor = "green"),
menuItem("User Journey", icon = icon("send", lib = "glyphicon"), tabName = "landing_page"
#,
#badgeLabel = "path", badgeColor = "orange"
)
# ,
# menuItem("valueBox test", tabName = "valueboxtest", icon = icon("thumbs-up", lib = "glyphicon"))
)
)
body <- dashboardBody(
tabItems(
tabItem(tabName = "social_location",
fluidRow(
column(
width = 5,
box(width = 12,
title = "DA webinar registrations - July 2020",
plotlyOutput("bulletgraph13", height = "120px")
)
),
column(width = 2,
br(),
br(),
valueBox(width = 12, goals$goal17Completions[5] + goals$goal13Completions[5], "Total signups (July)", icon = icon("list"))
),
column(width = 5,
box(width = 12,
title = "PSD webinar registrations - July 2020",
plotlyOutput("bulletgraph17", height = "120px")
)
)
),
fluidRow(
width = 12,
plotOutput("goals")
),
fluidRow(
column(width = 6,
br(),
plotOutput("social")
),
column(width = 6,
br(),
plotOutput("referrers")
)
),
),
tabItem(tabName = "landing_page",
fluidRow(
box(width = 12,
plotOutput("landing")
)
),
fluidRow(
box(width = 6,
plotOutput("landing_goals13")
),
box(width = 6,
plotOutput("landing_goals17")
)
)
)
# tabItem(tabName = "valueboxtest",
# valueBox(50*2, "valueBox test", icon = icon("list")))
)
)
ui <- dashboardPage(
dashboardHeader(title = "Website Engagement"),
sidebar,
body
)
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/spline-correlog.r
\name{print.spline.correlog}
\alias{print.spline.correlog}
\title{Print function for spline.correlog objects}
\usage{
\method{print}{spline.correlog}(x, ...)
}
\arguments{
\item{x}{an object of class "spline.correlog", usually, as a result of a call to \code{spline.correlog} or related).}
\item{\dots}{other arguments}
}
\value{
The function-call is printed to screen.
}
\description{
`print' method for class "spline.correlog".
}
\seealso{
\code{\link{spline.correlog}}
}
| /man/print.spline.correlog.Rd | no_license | objornstad/ncf | R | false | true | 571 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/spline-correlog.r
\name{print.spline.correlog}
\alias{print.spline.correlog}
\title{Print function for spline.correlog objects}
\usage{
\method{print}{spline.correlog}(x, ...)
}
\arguments{
\item{x}{an object of class "spline.correlog", usually, as a result of a call to \code{spline.correlog} or related).}
\item{\dots}{other arguments}
}
\value{
The function-call is printed to screen.
}
\description{
`print' method for class "spline.correlog".
}
\seealso{
\code{\link{spline.correlog}}
}
|
plot_case_studies <- function(cs1,
cs2,
performance_stats,
perf_summaries,
case_studies,
cs_name = "realistic",
lcomp_years = "max",
plot_variable = "f",
plot_period = "middle",
modelo_name) {
if (lcomp_years == "max"){
lyears <- max(performance_stats$prop_years_lcomp_data)
} else if (lcomp_years == "min"){
lyears <- min(performance_stats$prop_years_lcomp_data)
}
cs_summary <- performance_stats %>%
ungroup() %>%
filter(case_study == case_study,
prop_years_lcomp_data == lyears ,
variable == plot_variable,
model == cs1 | model == cs2)
# cs_summary$model <- fct_recode(as.factor(cs_summary$model), "Lengths Only" = cs1,
# "Lengths + Economics" = cs2 )
cs_summ_plot <- cs_summary %>%
gather(metric, value, recent_rmse, recent_median_bias) %>%
ggplot(aes(value, fill = model)) +
geom_vline(aes(xintercept = 0), linetype = 2) +
geom_density_ridges(aes(x = value, y = metric, fill = model),
alpha = 0.3,
color = "darkgrey",
show.legend = F) +
scale_y_discrete(labels = c("bias", "rmse"), position = "right", name = element_blank())+
theme(
plot.margin = unit(c(0, 0, 0, 1), units = "lines"),
axis.title.y = element_blank(),
axis.title.x = element_blank()
) +
theme(legend.title = element_blank(), axis.text.x = element_text(size = 10)) +
ggsci::scale_fill_aaas() +
labs(caption = "Calculated over last 5 years", title = "B")
lcomp_years <- case_studies %>%
filter(case_study == cs_name,
prop_years_lcomp_data == lyears,
period == plot_period)
lcomp_years <- lcomp_years$prepped_fishery[[1]]$scrooge_data$length_comps_years
cs <- perf_summaries %>%
ungroup() %>%
filter(case_study == cs_name,
prop_years_lcomp_data == lyears,
model == cs1 | model == cs2)
fill_vector <- rep("transparent", length (unique(cs$year[cs$variable == "f"])))
fill_vector[lcomp_years] <- "tomato"
fill_vector <- c(fill_vector, fill_vector)
cs_plot <- cs %>%
filter(variable == plot_variable) %>%
mutate(year = year - min(year) + 1) %>%
mutate(lcomp_year = year %in% lcomp_years) %>%
arrange(experiment) %>%
ggplot() +
geom_point(aes(year, observed), size = 3, alpha = 0.75, shape = 21, fill =fill_vector) +
geom_ribbon(aes(year, ymin = lower_90, ymax = upper_90, fill = model), alpha = 0.2, show.legend = FALSE) +
geom_line(aes(year,mean_predicted, color = model), show.legend = FALSE) +
labs(title = "A",x = "Year", y = "Fishing Mortality", caption = "Points = Data, Shaded = Predictions") +
ggsci::scale_color_aaas() +
ggsci::scale_fill_aaas() +
facet_wrap(~model, labeller = labeller(model = modelo_name)) +
theme(plot.margin = unit(c(0,0,0,0), units = "lines"),
panel.spacing = unit(0.1, units = "lines"),
strip.text = element_text(size = 14))
go_away <- ls()[!ls() %in% c("cs_plot","cs_summ_plot")]
rm(list = go_away)
out <- cs_plot + cs_summ_plot + plot_layout(nrow = 1, ncol = 2, widths = c(3,1))
}
| /functions/plot_case_studies.R | permissive | DanOvando/scrooge | R | false | false | 3,140 | r | plot_case_studies <- function(cs1,
cs2,
performance_stats,
perf_summaries,
case_studies,
cs_name = "realistic",
lcomp_years = "max",
plot_variable = "f",
plot_period = "middle",
modelo_name) {
if (lcomp_years == "max"){
lyears <- max(performance_stats$prop_years_lcomp_data)
} else if (lcomp_years == "min"){
lyears <- min(performance_stats$prop_years_lcomp_data)
}
cs_summary <- performance_stats %>%
ungroup() %>%
filter(case_study == case_study,
prop_years_lcomp_data == lyears ,
variable == plot_variable,
model == cs1 | model == cs2)
# cs_summary$model <- fct_recode(as.factor(cs_summary$model), "Lengths Only" = cs1,
# "Lengths + Economics" = cs2 )
cs_summ_plot <- cs_summary %>%
gather(metric, value, recent_rmse, recent_median_bias) %>%
ggplot(aes(value, fill = model)) +
geom_vline(aes(xintercept = 0), linetype = 2) +
geom_density_ridges(aes(x = value, y = metric, fill = model),
alpha = 0.3,
color = "darkgrey",
show.legend = F) +
scale_y_discrete(labels = c("bias", "rmse"), position = "right", name = element_blank())+
theme(
plot.margin = unit(c(0, 0, 0, 1), units = "lines"),
axis.title.y = element_blank(),
axis.title.x = element_blank()
) +
theme(legend.title = element_blank(), axis.text.x = element_text(size = 10)) +
ggsci::scale_fill_aaas() +
labs(caption = "Calculated over last 5 years", title = "B")
lcomp_years <- case_studies %>%
filter(case_study == cs_name,
prop_years_lcomp_data == lyears,
period == plot_period)
lcomp_years <- lcomp_years$prepped_fishery[[1]]$scrooge_data$length_comps_years
cs <- perf_summaries %>%
ungroup() %>%
filter(case_study == cs_name,
prop_years_lcomp_data == lyears,
model == cs1 | model == cs2)
fill_vector <- rep("transparent", length (unique(cs$year[cs$variable == "f"])))
fill_vector[lcomp_years] <- "tomato"
fill_vector <- c(fill_vector, fill_vector)
cs_plot <- cs %>%
filter(variable == plot_variable) %>%
mutate(year = year - min(year) + 1) %>%
mutate(lcomp_year = year %in% lcomp_years) %>%
arrange(experiment) %>%
ggplot() +
geom_point(aes(year, observed), size = 3, alpha = 0.75, shape = 21, fill =fill_vector) +
geom_ribbon(aes(year, ymin = lower_90, ymax = upper_90, fill = model), alpha = 0.2, show.legend = FALSE) +
geom_line(aes(year,mean_predicted, color = model), show.legend = FALSE) +
labs(title = "A",x = "Year", y = "Fishing Mortality", caption = "Points = Data, Shaded = Predictions") +
ggsci::scale_color_aaas() +
ggsci::scale_fill_aaas() +
facet_wrap(~model, labeller = labeller(model = modelo_name)) +
theme(plot.margin = unit(c(0,0,0,0), units = "lines"),
panel.spacing = unit(0.1, units = "lines"),
strip.text = element_text(size = 14))
go_away <- ls()[!ls() %in% c("cs_plot","cs_summ_plot")]
rm(list = go_away)
out <- cs_plot + cs_summ_plot + plot_layout(nrow = 1, ncol = 2, widths = c(3,1))
}
|
# For compatibility with 2.2.21
.get_course_path <- function(){
tryCatch(swirl:::swirl_courses_dir(),
error = function(c) {file.path(find.package("swirl"),"Courses")}
)
}
# Put initialization code in this file.
path_to_course <- file.path(.get_course_path(),
"R_102 - Getting_and_Cleaning_Data","ODBC")
| /R_102 - Getting_and_Cleaning_Data/ODBC/initLesson.R | no_license | ImprovementPathSystems/IPS_swirl_beta | R | false | false | 321 | r |
# For compatibility with 2.2.21
.get_course_path <- function(){
tryCatch(swirl:::swirl_courses_dir(),
error = function(c) {file.path(find.package("swirl"),"Courses")}
)
}
# Put initialization code in this file.
path_to_course <- file.path(.get_course_path(),
"R_102 - Getting_and_Cleaning_Data","ODBC")
|
#' Change PCA Table names
#'
#' @param table created with the \code{sas_prcomp_PCA_table_function} function.
#' @description First create a table from SAS output, then use this function to change the values of a couple of columns.
#'
#' @export
sas_PCA_table_names <- function(table) {
table %<>%
mutate(`Summary Statistic` = Var) %>%
mutate(
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="A1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="A2"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="B"),
""
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="C1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="C2"),
"Max."
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="C3"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="D1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="D2"),
"Max."
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="D3"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="E1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="E2"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="F"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="G"),
""
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="H1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="H2"),
"Max."
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="H3"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="Eigenvalue"),
""
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="Cumulative Proportion of Variance Explained"),
""
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="Rotated"),
""
)
) %>%
# mutate(
# `All Data PC 2` = replace(
# `All Data PC 2`,
# which(is.na(`All Data PC 2`)),
# ""
# )
# ) %>%
mutate(
Var = replace(
Var,
which(Var=="A1"),
"Daily Precipitation"
),
Var = replace(
Var,
which(Var=="A2"),
""
),
Var = replace(
Var,
which(Var=="B"),
"Percentage of Days with Rain"
),
Var = replace(
Var,
which(Var=="C1"),
"Number of Consecutive Days with Rain"
),
Var = replace(
Var,
which(Var=="C2"),
""
),
Var = replace(
Var,
which(Var=="C3"),
""
),
Var = replace(
Var,
which(Var=="D1"),
"Number of Consecutive Days without Rain"
),
Var = replace(
Var,
which(Var=="D2"),
""
),
Var = replace(
Var,
which(Var=="D3"),
""
),
Var = replace(
Var,
which(Var=="E1"),
"Maximum Temperature"
),
Var = replace(
Var,
which(Var=="E2"),
""
),
Var = replace(
Var,
which(Var=="F"),
"Mean Degree Day"
),
Var = replace(
Var,
which(Var=="G"),
"Percentage of Freezing Days"),
Var = replace(
Var,
which(Var=="H1"),
"Number of Cosecutive Days with Temperatures Below Freezing"
),
Var = replace(
Var,
which(Var=="H2"),
""
),
Var = replace(
Var,
which(Var=="H3"),
""
)
) %>%
select(Var, `Summary Statistic`, everything())
names(table)[names(table)=="Var"] <- "Variable"
return(table)
} | /R/SAS_PCA_table_names.R | no_license | ksauby/modresproc | R | false | false | 4,101 | r | #' Change PCA Table names
#'
#' @param table created with the \code{sas_prcomp_PCA_table_function} function.
#' @description First create a table from SAS output, then use this function to change the values of a couple of columns.
#'
#' @export
sas_PCA_table_names <- function(table) {
table %<>%
mutate(`Summary Statistic` = Var) %>%
mutate(
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="A1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="A2"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="B"),
""
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="C1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="C2"),
"Max."
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="C3"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="D1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="D2"),
"Max."
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="D3"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="E1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="E2"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="F"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="G"),
""
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="H1"),
"Mean"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="H2"),
"Max."
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="H3"),
"SD"
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="Eigenvalue"),
""
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="Cumulative Proportion of Variance Explained"),
""
),
`Summary Statistic` = replace(
`Summary Statistic`,
which(`Summary Statistic`=="Rotated"),
""
)
) %>%
# mutate(
# `All Data PC 2` = replace(
# `All Data PC 2`,
# which(is.na(`All Data PC 2`)),
# ""
# )
# ) %>%
mutate(
Var = replace(
Var,
which(Var=="A1"),
"Daily Precipitation"
),
Var = replace(
Var,
which(Var=="A2"),
""
),
Var = replace(
Var,
which(Var=="B"),
"Percentage of Days with Rain"
),
Var = replace(
Var,
which(Var=="C1"),
"Number of Consecutive Days with Rain"
),
Var = replace(
Var,
which(Var=="C2"),
""
),
Var = replace(
Var,
which(Var=="C3"),
""
),
Var = replace(
Var,
which(Var=="D1"),
"Number of Consecutive Days without Rain"
),
Var = replace(
Var,
which(Var=="D2"),
""
),
Var = replace(
Var,
which(Var=="D3"),
""
),
Var = replace(
Var,
which(Var=="E1"),
"Maximum Temperature"
),
Var = replace(
Var,
which(Var=="E2"),
""
),
Var = replace(
Var,
which(Var=="F"),
"Mean Degree Day"
),
Var = replace(
Var,
which(Var=="G"),
"Percentage of Freezing Days"),
Var = replace(
Var,
which(Var=="H1"),
"Number of Cosecutive Days with Temperatures Below Freezing"
),
Var = replace(
Var,
which(Var=="H2"),
""
),
Var = replace(
Var,
which(Var=="H3"),
""
)
) %>%
select(Var, `Summary Statistic`, everything())
names(table)[names(table)=="Var"] <- "Variable"
return(table)
} |
## Put comments here that give an overall description of what your
## functions do
## Write a short comment describing this function
makeCacheMatrix <- function(x = matrix()) {
## Initialize the inverse property
i <- NULL
## Method to set the matrix
set <- function( matrix ) {
m <<- matrix
i <<- NULL
}
## Method the get the matrix
get <- function() {
## Return the matrix
m
}
## Method to set the inverse of the matrix
setInverse <- function(inverse) {
i <<- inverse
}
## Method to get the inverse of the matrix
getInverse <- function() {
## Return the inverse property
i
}
## Return a list of the methods
list(set = set, get = get,
setInverse = setInverse,
getInverse = getInverse)
}
## Write a short comment describing this function
cacheSolve <- function(x, ...) {
## Return a matrix that is the inverse of 'x'
m <- x$getInverse()
## Just return the inverse if its already set
if( !is.null(m) ) {
message("getting cached data")
return(m)
}
## Get the matrix from our object
data <- x$get()
## Calculate the inverse using matrix multiplication
m <- solve(data) %*% data
## Set the inverse to the object
x$setInverse(m)
## Return the matrix
m
}
| /cachematrix.R | no_license | manavjain174/ProgrammingAssignment2 | R | false | false | 1,362 | r | ## Put comments here that give an overall description of what your
## functions do
## Write a short comment describing this function
makeCacheMatrix <- function(x = matrix()) {
## Initialize the inverse property
i <- NULL
## Method to set the matrix
set <- function( matrix ) {
m <<- matrix
i <<- NULL
}
## Method the get the matrix
get <- function() {
## Return the matrix
m
}
## Method to set the inverse of the matrix
setInverse <- function(inverse) {
i <<- inverse
}
## Method to get the inverse of the matrix
getInverse <- function() {
## Return the inverse property
i
}
## Return a list of the methods
list(set = set, get = get,
setInverse = setInverse,
getInverse = getInverse)
}
## Write a short comment describing this function
cacheSolve <- function(x, ...) {
## Return a matrix that is the inverse of 'x'
m <- x$getInverse()
## Just return the inverse if its already set
if( !is.null(m) ) {
message("getting cached data")
return(m)
}
## Get the matrix from our object
data <- x$get()
## Calculate the inverse using matrix multiplication
m <- solve(data) %*% data
## Set the inverse to the object
x$setInverse(m)
## Return the matrix
m
}
|
testlist <- list(testX = c(191493125665849920, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), trainX = structure(c(1.78844646178735e+212, 1.93075223605916e+156, 121373.193669204, 1.26689771433298e+26, 2.46020195254853e+129, 8.54794497535107e-83, 2.61907806894971e-213, 1.5105425626729e+200, 6.51877713351675e+25, 4.40467528702727e-93, 7.6427933587945, 34208333744.1307, 1.6400690920442e-111, 3.9769673154778e-304, 4.76127371594362e-307, 8.63819952335095e+122, 1.18662128550178e-59, 1128.83285802937, 3.80368040657947e-72, 1.21321365773924e-195, 9.69744674150153e-268, 8.98899319496613e+272, 7.63669788330223e+285, 3.85830749537493e+266, 2.65348875902107e+136, 8.14965241967603e+92, 2.59677146539475e-173, 1.55228780425777e-91, 8.25550184376779e+105, 1.18572662524891e+134, 1.04113208597565e+183, 1.01971211553913e-259, 1.23680594512923e-165, 5.24757023065221e+62, 3.41816623041351e-96 ), .Dim = c(5L, 7L)))
result <- do.call(dann:::calc_distance_C,testlist)
str(result) | /dann/inst/testfiles/calc_distance_C/AFL_calc_distance_C/calc_distance_C_valgrind_files/1609867112-test.R | no_license | akhikolla/updated-only-Issues | R | false | false | 1,199 | r | testlist <- list(testX = c(191493125665849920, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), trainX = structure(c(1.78844646178735e+212, 1.93075223605916e+156, 121373.193669204, 1.26689771433298e+26, 2.46020195254853e+129, 8.54794497535107e-83, 2.61907806894971e-213, 1.5105425626729e+200, 6.51877713351675e+25, 4.40467528702727e-93, 7.6427933587945, 34208333744.1307, 1.6400690920442e-111, 3.9769673154778e-304, 4.76127371594362e-307, 8.63819952335095e+122, 1.18662128550178e-59, 1128.83285802937, 3.80368040657947e-72, 1.21321365773924e-195, 9.69744674150153e-268, 8.98899319496613e+272, 7.63669788330223e+285, 3.85830749537493e+266, 2.65348875902107e+136, 8.14965241967603e+92, 2.59677146539475e-173, 1.55228780425777e-91, 8.25550184376779e+105, 1.18572662524891e+134, 1.04113208597565e+183, 1.01971211553913e-259, 1.23680594512923e-165, 5.24757023065221e+62, 3.41816623041351e-96 ), .Dim = c(5L, 7L)))
result <- do.call(dann:::calc_distance_C,testlist)
str(result) |
library(here)
library(keras)
library(tidytext)
library(tidyverse)
set.seed(55)
tweets <- read_csv(here("tweets", "truthdata_allrounds.csv")) %>%
mutate(accident = NA) %>%
mutate(accident = replace(accident, accident_truth == "No", 0)) %>%
mutate(accident = replace(accident, accident_truth == "Yes", 1)) %>%
select("tweet", "accident") %>% na.omit
table(tweets$accident)
# pre-process text
text_process <- function(text, stopwords = FALSE, lower = FALSE) {
# remove time stamp and other numbers
#text_data <- gsub("[[:digit:]]*|\\r", "", text)
cat_text <- paste(text, collapse=" ")
text_sequence <- unique(text_to_word_sequence(cat_text, split = " ", lower = lower))
if(stopwords) {
text_index <- tibble(word = text_sequence, index = 1:length(text_sequence)) %>%
anti_join(get_stopwords(), by = "word")
} else {
text_index <- tibble(word = text_sequence, index = 1:length(text_sequence))
}
text_index
}
# create a word index to match text to integers
word_index <- text_process(tweets$tweet)
# make integer sequences of tweet text
make_index_list <- function(x) {
data_index <- c()
for (i in 1:length(x)) {
text <- x[[i]]
x_seq <- text_to_word_sequence(text, split = " ", lower = FALSE)
x_int <- c()
for(n in x_seq) {
int <- word_index$index[word_index$word %in% n]
x_int <- c(x_int, int)
x_int
}
data_index[[i]] <- x_int
}
data_index
}
# hot encode integer sequences
vectorize_sequences <- function(sequences, dimension = nrow(word_index)) {
results <- matrix(0, nrow = length(sequences), ncol = dimension)
for (i in 1:length(sequences)) {
results[i, sequences[[i]]] <- 1
}
results
}
# separate tweet data into training and testing sets
index2train <- sample(1:nrow(tweets), nrow(tweets)/2)
# data
train_data <- tweets$tweet[index2train]
test_data <- tweets$tweet[-index2train]
# labels
y_train <- tweets$accident[index2train]
y_test <- tweets$accident[-index2train]
# convert words to integers based on their place in the dictionary
test_index <- make_index_list(test_data)
train_index <- make_index_list(train_data)
x_train <- vectorize_sequences(train_index)
x_test <- vectorize_sequences(test_index)
#val_indices <- sample(1:nrow(x_train), 500)
val_indices <- 1:500
x_val <- x_train[val_indices,]
partial_x_train <- x_train[-val_indices,]
y_val <- y_train[val_indices]
partial_y_train <- y_train[-val_indices]
model <- keras_model_sequential() %>%
layer_dense(units = 16, activation = "relu", input_shape = nrow(word_index)) %>%
layer_dense(units = 16, activation = "relu") %>%
layer_dense(units = 1, activation = "sigmoid")
model %>% compile(
optimizer = "rmsprop",
loss = "binary_crossentropy",
metrics = c("accuracy")
)
history <- model %>% fit(
partial_x_train,
partial_y_train,
epochs = 50,
batch_size = 300,
validation_data = list(x_val, y_val)
)
plot(history)
# retrain model with peak epochs from validation
model <- keras_model_sequential() %>%
layer_dense(units = 16, activation = "relu", input_shape = nrow(word_index)) %>%
layer_dense(units = 16, activation = "relu") %>%
layer_dense(units = 1, activation = "sigmoid")
model %>% compile(
optimizer = "rmsprop",
loss = "binary_crossentropy",
metrics = c("accuracy")
)
model %>% fit(
x_train,
y_train,
epochs = 32,
batch_size = 300
)
results <- model %>% evaluate(x_test, y_test)
results
predictions <- model %>% predict(x_test)
predictions_check <- tibble(tweet = test_data, accident = y_test, prediction = predictions[,1]) %>%
mutate(predict_accident = 0) %>%
mutate(predict_accident = replace(predict_accident, prediction > .70, 1)) %>%
mutate(correct = case_when(accident == predict_accident ~ 1)) %>%
mutate(correct = replace(correct, is.na(correct), 0))
table(predictions_check$predict_accident)
table(predictions_check$accident)
mean(predictions_check$correct)
### Function for predicting new Tweets
predict_accident <- function(tweet) {
tweet_index <- make_index_list(tweet)
tweet_vectorized <- vectorize_sequences(tweet_index)
prediction <- model %>% predict(tweet_vectorized)
if(prediction > .80) {
list("This Tweet describes an accident.", tweet, prediction)
} else {
list("This Tweet does not describe an accident.", tweet, prediction)
}
}
predict_accident("22:36 @KeNHAKenya the pothole is still there! (as you approach sigona from Limuru)Those are 4 cars with punctures! At this hour- and in the rain . For how long will Kenyans suffer?? @JamesMacharia_ via @bnyenjeri")
predict_accident("heavy traffic between Roysambu zimmerman. alternative routes via")
predict_accident("21:33 Ruaka route is really bad!. No cops to Assist control traffic. Sad!! via @FaithMunyasi")
predict_accident("18:25 hit and run on Waiyaki Way causing snarl up. https://twitter.com/aqueerian/status/992787431157129217/photo/1pic.twitter.com/aaKMgmf8Zi via @aqueerian")
predict_accident("18:24 Has anyone ever received these monies police claim? Witnesses come forward,,me i know ukiwaambia unamjua ndo utajua hujui via @Mkulima18")
predict_accident("JAPAN: Bus drivers in Okayama have gone on strike, by continuing to drive their routes while refusing to take fares from passengers.")
predict_accident("16:57 Avoid Outering Rd from Taj Mall towards Donholm.Traffic bumper to bumper, via @CaroleOO16")
predict_accident("16:55 I passed their a few hours ago and was astonished by how drivers can be...parking vehicles on the road to go and peep at a crashed vehicle causing huge traffic snarl up ..... Every parked vehicle needed to be booked for obstruction via @MateMurithi")
predict_accident("16:27 Stuck truck at Outering exit juu to Doni caltex frm pipeline creating jam kubwa via @simwa_")
predict_accident("15:38 Very heavy traffic in Juja on your way to Thika. Still trying to recover the lorry that plunged into Ndarugo river. via @ThikaTowntoday")
predict_accident("12:52 Alert: Accident involving motorbike and private car at the Bomas of Kenya near \"KWS\" College on Langata Road. Exercise caution. + No police on sight. https://twitter.com/GrayMarwa/status/992697224583897088 β¦ via @GrayMarwa")
predict_accident("12:28 Snarl up on Langata Road caused by accident involving a motorcycle & motor vehicle after Bomas before the National Park. via @Jaxon09")
predict_accident("12:18 From South c to Uhuru High way towards CBD moving smoothlyπππππππ Mombasa Road via @Eng_Onsoti")
predict_accident("12:09 @NairobiAlerts @digitalmkenya Our laws are funny though,, the prosecutor had to prove the pastor intended to kill the said victim. Otherwise, it wil be treated as an accident and the best the family can do is to follow https://www.ma3route.com/update/703679 via @MatthewsMbuthia")
predict_accident("12:02 @alfredootieno He promised a change in 100 days. What we have seen is painful.
Water shortage is even beyond words can describe. Traffic has become the order of the day.
What kind of patience are you talking about? via @AderaJacob")
predict_accident("11:21 avoid route between muthaiga mini mkt and parklands. Bumper to bumper traffic as usual. via @Its_Sheshii")
predict_accident("10:47 @Dj_Amar_B Riverside drive has become a construction site. I assume someone is blasting some site. via @Pagmainc")
predict_accident("10:32 Speed trap as you approach Kinoo.. via @njauwamahugu")
predict_accident("10:29 @KenyaPower_Care A pole has fallen/leaning on a vehicle along Gaberone rd off Luthuli avenue via @kimaniq")
predict_accident("Fatal accident at Kiwanja on Nothern bypass cleared.huge crowd on the road side. Boda boda rider taken to hospital.the deceased was a pillion passenger @Kiss100kenya @PRSA_Roadsafety @NPSOfficial_KE @KURAroads @RadioJamboKenya @inooroke @CapitalFMKenya @2Fmke @K24Tv @ntsa_kenya
")
predict_accident("10:11 @NPSOfficial_KE corruption thriving along Msa road just after machakos towards konza. cops in a highway xtrail patrol car via @cymohnganga")
# remove hashtags and some stop words (in a)
predict_accident("corruption thriving along Msa road just after machakos towards konza. cops highway xtrail patrol car via")
predict_accident("09:39 Via @MwendeCharles Nothing beats a Subaru. The way these machines negotiate sharp bends at 150Km/hr! Damn! @NairobiAlerts via @digitalmkenya")
| /scripts/classify_32_300_90.R | no_license | robtenorio/classify-accidents | R | false | false | 8,390 | r | library(here)
library(keras)
library(tidytext)
library(tidyverse)
set.seed(55)
tweets <- read_csv(here("tweets", "truthdata_allrounds.csv")) %>%
mutate(accident = NA) %>%
mutate(accident = replace(accident, accident_truth == "No", 0)) %>%
mutate(accident = replace(accident, accident_truth == "Yes", 1)) %>%
select("tweet", "accident") %>% na.omit
table(tweets$accident)
# pre-process text
text_process <- function(text, stopwords = FALSE, lower = FALSE) {
# remove time stamp and other numbers
#text_data <- gsub("[[:digit:]]*|\\r", "", text)
cat_text <- paste(text, collapse=" ")
text_sequence <- unique(text_to_word_sequence(cat_text, split = " ", lower = lower))
if(stopwords) {
text_index <- tibble(word = text_sequence, index = 1:length(text_sequence)) %>%
anti_join(get_stopwords(), by = "word")
} else {
text_index <- tibble(word = text_sequence, index = 1:length(text_sequence))
}
text_index
}
# create a word index to match text to integers
word_index <- text_process(tweets$tweet)
# make integer sequences of tweet text
make_index_list <- function(x) {
data_index <- c()
for (i in 1:length(x)) {
text <- x[[i]]
x_seq <- text_to_word_sequence(text, split = " ", lower = FALSE)
x_int <- c()
for(n in x_seq) {
int <- word_index$index[word_index$word %in% n]
x_int <- c(x_int, int)
x_int
}
data_index[[i]] <- x_int
}
data_index
}
# hot encode integer sequences
vectorize_sequences <- function(sequences, dimension = nrow(word_index)) {
results <- matrix(0, nrow = length(sequences), ncol = dimension)
for (i in 1:length(sequences)) {
results[i, sequences[[i]]] <- 1
}
results
}
# separate tweet data into training and testing sets
index2train <- sample(1:nrow(tweets), nrow(tweets)/2)
# data
train_data <- tweets$tweet[index2train]
test_data <- tweets$tweet[-index2train]
# labels
y_train <- tweets$accident[index2train]
y_test <- tweets$accident[-index2train]
# convert words to integers based on their place in the dictionary
test_index <- make_index_list(test_data)
train_index <- make_index_list(train_data)
x_train <- vectorize_sequences(train_index)
x_test <- vectorize_sequences(test_index)
#val_indices <- sample(1:nrow(x_train), 500)
val_indices <- 1:500
x_val <- x_train[val_indices,]
partial_x_train <- x_train[-val_indices,]
y_val <- y_train[val_indices]
partial_y_train <- y_train[-val_indices]
model <- keras_model_sequential() %>%
layer_dense(units = 16, activation = "relu", input_shape = nrow(word_index)) %>%
layer_dense(units = 16, activation = "relu") %>%
layer_dense(units = 1, activation = "sigmoid")
model %>% compile(
optimizer = "rmsprop",
loss = "binary_crossentropy",
metrics = c("accuracy")
)
history <- model %>% fit(
partial_x_train,
partial_y_train,
epochs = 50,
batch_size = 300,
validation_data = list(x_val, y_val)
)
plot(history)
# retrain model with peak epochs from validation
model <- keras_model_sequential() %>%
layer_dense(units = 16, activation = "relu", input_shape = nrow(word_index)) %>%
layer_dense(units = 16, activation = "relu") %>%
layer_dense(units = 1, activation = "sigmoid")
model %>% compile(
optimizer = "rmsprop",
loss = "binary_crossentropy",
metrics = c("accuracy")
)
model %>% fit(
x_train,
y_train,
epochs = 32,
batch_size = 300
)
results <- model %>% evaluate(x_test, y_test)
results
predictions <- model %>% predict(x_test)
predictions_check <- tibble(tweet = test_data, accident = y_test, prediction = predictions[,1]) %>%
mutate(predict_accident = 0) %>%
mutate(predict_accident = replace(predict_accident, prediction > .70, 1)) %>%
mutate(correct = case_when(accident == predict_accident ~ 1)) %>%
mutate(correct = replace(correct, is.na(correct), 0))
table(predictions_check$predict_accident)
table(predictions_check$accident)
mean(predictions_check$correct)
### Function for predicting new Tweets
predict_accident <- function(tweet) {
tweet_index <- make_index_list(tweet)
tweet_vectorized <- vectorize_sequences(tweet_index)
prediction <- model %>% predict(tweet_vectorized)
if(prediction > .80) {
list("This Tweet describes an accident.", tweet, prediction)
} else {
list("This Tweet does not describe an accident.", tweet, prediction)
}
}
predict_accident("22:36 @KeNHAKenya the pothole is still there! (as you approach sigona from Limuru)Those are 4 cars with punctures! At this hour- and in the rain . For how long will Kenyans suffer?? @JamesMacharia_ via @bnyenjeri")
predict_accident("heavy traffic between Roysambu zimmerman. alternative routes via")
predict_accident("21:33 Ruaka route is really bad!. No cops to Assist control traffic. Sad!! via @FaithMunyasi")
predict_accident("18:25 hit and run on Waiyaki Way causing snarl up. https://twitter.com/aqueerian/status/992787431157129217/photo/1pic.twitter.com/aaKMgmf8Zi via @aqueerian")
predict_accident("18:24 Has anyone ever received these monies police claim? Witnesses come forward,,me i know ukiwaambia unamjua ndo utajua hujui via @Mkulima18")
predict_accident("JAPAN: Bus drivers in Okayama have gone on strike, by continuing to drive their routes while refusing to take fares from passengers.")
predict_accident("16:57 Avoid Outering Rd from Taj Mall towards Donholm.Traffic bumper to bumper, via @CaroleOO16")
predict_accident("16:55 I passed their a few hours ago and was astonished by how drivers can be...parking vehicles on the road to go and peep at a crashed vehicle causing huge traffic snarl up ..... Every parked vehicle needed to be booked for obstruction via @MateMurithi")
predict_accident("16:27 Stuck truck at Outering exit juu to Doni caltex frm pipeline creating jam kubwa via @simwa_")
predict_accident("15:38 Very heavy traffic in Juja on your way to Thika. Still trying to recover the lorry that plunged into Ndarugo river. via @ThikaTowntoday")
predict_accident("12:52 Alert: Accident involving motorbike and private car at the Bomas of Kenya near \"KWS\" College on Langata Road. Exercise caution. + No police on sight. https://twitter.com/GrayMarwa/status/992697224583897088 β¦ via @GrayMarwa")
predict_accident("12:28 Snarl up on Langata Road caused by accident involving a motorcycle & motor vehicle after Bomas before the National Park. via @Jaxon09")
predict_accident("12:18 From South c to Uhuru High way towards CBD moving smoothlyπππππππ Mombasa Road via @Eng_Onsoti")
predict_accident("12:09 @NairobiAlerts @digitalmkenya Our laws are funny though,, the prosecutor had to prove the pastor intended to kill the said victim. Otherwise, it wil be treated as an accident and the best the family can do is to follow https://www.ma3route.com/update/703679 via @MatthewsMbuthia")
predict_accident("12:02 @alfredootieno He promised a change in 100 days. What we have seen is painful.
Water shortage is even beyond words can describe. Traffic has become the order of the day.
What kind of patience are you talking about? via @AderaJacob")
predict_accident("11:21 avoid route between muthaiga mini mkt and parklands. Bumper to bumper traffic as usual. via @Its_Sheshii")
predict_accident("10:47 @Dj_Amar_B Riverside drive has become a construction site. I assume someone is blasting some site. via @Pagmainc")
predict_accident("10:32 Speed trap as you approach Kinoo.. via @njauwamahugu")
predict_accident("10:29 @KenyaPower_Care A pole has fallen/leaning on a vehicle along Gaberone rd off Luthuli avenue via @kimaniq")
predict_accident("Fatal accident at Kiwanja on Nothern bypass cleared.huge crowd on the road side. Boda boda rider taken to hospital.the deceased was a pillion passenger @Kiss100kenya @PRSA_Roadsafety @NPSOfficial_KE @KURAroads @RadioJamboKenya @inooroke @CapitalFMKenya @2Fmke @K24Tv @ntsa_kenya
")
predict_accident("10:11 @NPSOfficial_KE corruption thriving along Msa road just after machakos towards konza. cops in a highway xtrail patrol car via @cymohnganga")
# remove hashtags and some stop words (in a)
predict_accident("corruption thriving along Msa road just after machakos towards konza. cops highway xtrail patrol car via")
predict_accident("09:39 Via @MwendeCharles Nothing beats a Subaru. The way these machines negotiate sharp bends at 150Km/hr! Damn! @NairobiAlerts via @digitalmkenya")
|
context("list")
skip_on_cran()
skip_if_not_installed("modeltests")
library(modeltests)
test_that("not all lists can be tidied", {
nl <- list(a = NULL)
expect_error(tidy(nl), "No tidy method recognized for this list.")
expect_error(glance(nl), "No glance method recognized for this list.")
})
| /tests/testthat/test-list.R | permissive | tidymodels/broom | R | false | false | 302 | r | context("list")
skip_on_cran()
skip_if_not_installed("modeltests")
library(modeltests)
test_that("not all lists can be tidied", {
nl <- list(a = NULL)
expect_error(tidy(nl), "No tidy method recognized for this list.")
expect_error(glance(nl), "No glance method recognized for this list.")
})
|
\name{edbGetActiveParticipants}
\alias{edbGetActiveParticipants}
\title{Returns dataframe with info about all active participants}
\description{
Returns dataframe with info about all active participants
}
\usage{edbGetActiveParticipants(d)}
\arguments{
\item{d}{cleaned enrollment database as dataframe}
}
\value{Returns a dataframe}
\author{John J. Curtin \email{jjcurtin@wisc.edu}} | /man/edbGetActiveParticipants.Rd | no_license | jjcurtin/StudySupport | R | false | false | 385 | rd | \name{edbGetActiveParticipants}
\alias{edbGetActiveParticipants}
\title{Returns dataframe with info about all active participants}
\description{
Returns dataframe with info about all active participants
}
\usage{edbGetActiveParticipants(d)}
\arguments{
\item{d}{cleaned enrollment database as dataframe}
}
\value{Returns a dataframe}
\author{John J. Curtin \email{jjcurtin@wisc.edu}} |
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/export.r
\name{CreateCsv}
\alias{CreateCsv}
\title{Create CSV}
\usage{
CreateCsv(ediData, filePath)
}
\arguments{
\item{ediData}{The EDI data in format of data frame or ffdf containing the data to be inserted.}
\item{filePath}{path and file name where CSV is written}
}
\description{
Create CSV
}
\details{
Generate CSV with snake-case columns
}
| /man/CreateCsv.Rd | permissive | parkyijoo/EdiToOmop | R | false | true | 425 | rd | % Generated by roxygen2: do not edit by hand
% Please edit documentation in R/export.r
\name{CreateCsv}
\alias{CreateCsv}
\title{Create CSV}
\usage{
CreateCsv(ediData, filePath)
}
\arguments{
\item{ediData}{The EDI data in format of data frame or ffdf containing the data to be inserted.}
\item{filePath}{path and file name where CSV is written}
}
\description{
Create CSV
}
\details{
Generate CSV with snake-case columns
}
|
library(CaDENCE)
### Name: cadence.predict
### Title: Predict conditional distribution parameters from a fitted CDEN
### model
### Aliases: cadence.predict cadence.evaluate
### ** Examples
data(FraserSediment)
lnorm.distribution.fixed <- list(density.fcn = dlnorm,
parameters = c("meanlog", "sdlog"),
parameters.fixed = "sdlog",
output.fcns = c(identity, exp))
fit <- cadence.fit(x = FraserSediment$x.1970.1976,
y = FraserSediment$y.1970.1976,
hidden.fcn = identity, maxit.Nelder = 100,
trace.Nelder = 1, trace = 1,
distribution = lnorm.distribution.fixed)
pred <- cadence.predict(x = FraserSediment$x.1977.1979, fit = fit)
matplot(pred, type = "l")
| /data/genthat_extracted_code/CaDENCE/examples/cadence.predict.Rd.R | no_license | surayaaramli/typeRrh | R | false | false | 882 | r | library(CaDENCE)
### Name: cadence.predict
### Title: Predict conditional distribution parameters from a fitted CDEN
### model
### Aliases: cadence.predict cadence.evaluate
### ** Examples
data(FraserSediment)
lnorm.distribution.fixed <- list(density.fcn = dlnorm,
parameters = c("meanlog", "sdlog"),
parameters.fixed = "sdlog",
output.fcns = c(identity, exp))
fit <- cadence.fit(x = FraserSediment$x.1970.1976,
y = FraserSediment$y.1970.1976,
hidden.fcn = identity, maxit.Nelder = 100,
trace.Nelder = 1, trace = 1,
distribution = lnorm.distribution.fixed)
pred <- cadence.predict(x = FraserSediment$x.1977.1979, fit = fit)
matplot(pred, type = "l")
|
\name{b.com.est}
\alias{b.com.est}
\title{ Common slope estimation }
\description{
Estimates a common slope from bivariate data for several independent groups.
Called by \code{\link{slope.com}}.
}
\usage{
b.com.est(z, n, method, lambda = 1, res.df)
}
\arguments{
\item{z}{ Variances and covariances of each group. }
\item{n}{ Sample sizes of each group. }
\item{method}{ See \code{\link{slope.com}} for details. }
\item{lambda}{ Error variance ration (implied by choice of method). }
\item{res.df}{ Residual degrees of freedom, for each group. }
}
\author{ Warton, D. \email{David.Warton@unsw.edu.au} and J. Ormerod }
\seealso{ \code{\link{slope.com}} }
\keyword{ internal }
| /man/b.com.est.Rd | no_license | cran/smatr | R | false | false | 714 | rd | \name{b.com.est}
\alias{b.com.est}
\title{ Common slope estimation }
\description{
Estimates a common slope from bivariate data for several independent groups.
Called by \code{\link{slope.com}}.
}
\usage{
b.com.est(z, n, method, lambda = 1, res.df)
}
\arguments{
\item{z}{ Variances and covariances of each group. }
\item{n}{ Sample sizes of each group. }
\item{method}{ See \code{\link{slope.com}} for details. }
\item{lambda}{ Error variance ration (implied by choice of method). }
\item{res.df}{ Residual degrees of freedom, for each group. }
}
\author{ Warton, D. \email{David.Warton@unsw.edu.au} and J. Ormerod }
\seealso{ \code{\link{slope.com}} }
\keyword{ internal }
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.