Spaces:
Sleeping
Sleeping
File size: 2,856 Bytes
280b65c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 | # Stage 3: append transit to the reference and recompute percentiles.
#
# IMPORTANT: BAS is the percentile of accessible richness WITHIN the same mode
# and the same travel time. Adding transit therefore creates a new percentile
# pool of its own and does NOT move walking, cycling or driving by a single
# point. That is the property that makes "relative percentages" comparable
# across modes: each mode is ranked against itself, so a 70th-percentile
# transit origin and a 70th-percentile walking origin both mean "better than
# 70% of the city, for someone travelling that way".
suppressMessages({library(data.table)})
setwd("~/Downloads/Biodiversity_Access_Score")
REF <- "outputs/results/isochrone_basi_100m.rds"
pctile <- function(x) { # identical to the analysis script
ok <- is.finite(x); o <- rep(NA_real_, length(x))
if (!any(ok)) return(o)
o[ok] <- 100*(rank(x[ok], ties.method="average") - 0.5)/sum(ok); o
}
ref <- as.data.table(readRDS(REF))
old_modes <- sort(unique(ref$mode))
if ("transit" %in% old_modes) { ref <- ref[mode != "transit"]; cat("replacing existing transit rows\n") }
TR <- as.data.table(readRDS("outputs/results/transit/transit_reference_rows.rds"))
# snapshot walking BAS so we can prove the other modes are untouched
before <- ref[mode=="walking" & time_min==15, .(cell_id, old = pctile(n_species_accessed))]
for (cl in setdiff(names(ref), names(TR))) TR[[cl]] <- NA
TR <- TR[, names(ref), with=FALSE]
full <- rbind(ref, TR)
# recompute every percentile column within mode x time
pcols <- grep("^pctile_", names(full), value=TRUE)
for (pc in pcols) {
src <- sub("^pctile_", "", pc)
if (!src %in% names(full)) next
if (src == "sf_ej_score") { # inverted: high burden = low percentile
full[, (pc) := 100 - pctile(get(src)), by=.(mode, time_min)]
} else if (src %in% c("nearest_stop_m","calenviro_ci_score","calenviro_traffic_pct")) {
full[, (pc) := 100 - pctile(get(src)), by=.(mode, time_min)]
} else {
full[, (pc) := pctile(get(src)), by=.(mode, time_min)]
}
}
after <- full[mode=="walking" & time_min==15, .(cell_id, new = pctile_n_species_accessed)]
chk <- merge(before, after, by="cell_id")
cat(sprintf("\nwalking-15 BAS unchanged by adding transit: max shift %.6f points\n",
max(abs(chk$old - chk$new), na.rm=TRUE)))
cat("\n=== reference now holds ===\n")
print(full[, .(origins=.N, median_species=median(n_species_accessed, na.rm=TRUE),
median_BAS=round(median(pctile_n_species_accessed, na.rm=TRUE),1)),
by=.(mode, time_min)][order(mode, time_min)])
bk <- sub("\\.rds$", "_pre_transit.rds", REF)
if (!file.exists(bk)) file.copy(REF, bk)
saveRDS(full, REF)
cat(sprintf("\nWROTE %s (%s rows, %d modes). Backup: %s\n",
REF, format(nrow(full), big.mark=","), length(unique(full$mode)), basename(bk)))
|