Update app.R

app.R

@@ -163,15 +163,86 @@ scale_TimStuff <- function(raw_score, model_mean, model_sd) {
   result <- 50 - (scaled_score * 10)
   return(result)
 }
+calculate_primary <- function(data){
+  data <- data %>%
+    # Group by pitch_name, Pitcher Name, Pitcher Id, and date
+    group_by(pitch_name, `Pitcher Name`, `Pitcher ID`, date) %>%
+    
+    # Count occurrences and calculate average start_speed, IVB, and HB for each group
+    mutate(
+      pitch_count = n(),
+      avg_start_speed = mean(start_speed, na.rm = TRUE),
+      avg_IVB = mean(IVB, na.rm = TRUE),
+      avg_HB = mean(HB, na.rm = TRUE)
+    ) %>%
+    
+    ungroup() %>%
+    # Group by Pitcher Name, Pitcher Id, and date
+    group_by(`Pitcher Name`, `Pitcher ID`, date) %>%
+    
+    # Add a column to identify the highest occurrence
+    mutate(
+      is_highest_occurrence = case_when(
+        pitch_count == max(pitch_count) ~ 1,
+        TRUE ~ 0
+      )
+    ) %>%
+    
+    # If there's a tie, use avg_start_speed as a tiebreaker
+    mutate(
+      is_highest_occurrence = case_when(
+        is_highest_occurrence == 1 & pitch_count == max(pitch_count[is_highest_occurrence == 1]) & 
+          avg_start_speed == max(avg_start_speed[is_highest_occurrence == 1]) ~ 1,
+        TRUE ~ 0
+      )
+    ) %>%
+    
+    # Calculate primary pitch metrics
+    mutate(
+      primary_speed = avg_start_speed[is_highest_occurrence == 1][1],
+      primary_IVB = avg_IVB[is_highest_occurrence == 1][1],
+      primary_HB = avg_HB[is_highest_occurrence == 1][1]
+    ) %>%
+    
+    # Ungroup to remove grouping structure
+    ungroup()
+}
 
+calculate_timstuff <- function(game) {
+  game <- calculate_primary(game)
+  game <- game %>%
+    mutate(VAA = calculate_VAA(vz0, ay, az, vy0, y0),
+           EAA = calculate_EAA(extension),
+           SADiff = calculate_SADiff(pfxX, pfxZ, spinDirection),
+           team_fielding_id = ifelse(description %in% c("Called Strike", "Swinging Strike", "Swinging Strike (Blocked)"), 1, 0),
+           swing = ifelse(description %in% c("Foul", "Foul Pitchout", "In play, no out", "In play, out(s)", "In play, run(s)", "Swinging Strike", "Swinging Strike (Blocked)", "Foul Tip"), 1, 0),
+           is_strike_swinging = ifelse(is_strike_swinging, 1, 0),
+           Pitch = pitch_name,
+           ishandL = ifelse(phand == "L",1,0))
+  # game <- calculate_primary(game)
+  
+  feature_vars <- c("ishandL","start_speed", "IVB", "HB", "EAA", "x0", "z0", "spin_rate","SADiff","primary_speed","primary_IVB","primary_HB")
+  complete_rows <- complete.cases(game[, feature_vars])
+  game_complete <- game[complete_rows, ]
+  game_na <- game[!complete_rows,]
+  game_na$TimStuff <- NA
+  
+  rhp <- game_complete
+  
+  # rhp <- game_complete[game_complete$ishandL == 0]
+  # 
+  # lhp$TimStuff <- scale_TimStuff(predict(model, as.matrix(cbind(lhp$ishandL,lhp$start_speed, lhp$IVB, lhp$HB, lhp$EAA, lhp$x0, lhp$z0, lhp$spin_rate, lhp$SADiff,lhp$primary_speed,lhp$primary_IVB,lhp$primary_HB))), -0.00249975,  0.007566558)
+  
+  rhp$TimStuff <- scale_TimStuff(predict(model, as.matrix(cbind(rhp$ishandL,rhp$start_speed, rhp$IVB, rhp$HB, rhp$EAA, rhp$x0, rhp$z0, rhp$spin_rate, rhp$SADiff,rhp$primary_speed,rhp$primary_IVB,rhp$primary_HB))), -0.00249975,  0.007566558)
+  
+  game_complete <- rbind(rhp,game_na)
+  return(game_complete)
+}
 summary_table <- function(game) {
   rows <- nrow(game)
+  game <- calculate_primary(game)
+  game <- calculate_TimStuff(game)
   sumtable <- game %>%
-    mutate(
-      VAA = calculate_VAA(vz0, ay,az, vy0, y0),
-      EAA = calculate_EAA(extension),
-      SADiff = calculate_SADiff(pfxX, pfxZ,spinDirection)
-    ) %>%
     mutate(team_fielding_id = ifelse(description == "Called Strike" |
                                        description == "Swinging Strike" |
                                        description == "Swinging Strike (Blocked)", 1, 0)) %>%
@@ -186,45 +257,6 @@ summary_table <- function(game) {
     mutate(is_strike_swinging = ifelse(is_strike_swinging == TRUE, 1, 0)) %>%
     mutate(Pitch = pitch_name) %>%
     rowwise() %>%
-    mutate(TimStuff = if (phand == 'L') {
-      case_when(
-        Pitch %in% c("Four-Seam Fastball", "Sinker", "Cutter", "Fastball") ~ 
-          ifelse(any(is.na(c(start_speed, IVB, -HB, EAA, -x0, z0, spin_rate, SADiff))), 
-                 NA_real_, 
-                 scale_TimStuff(predict(FB, matrix(c(start_speed, IVB, -HB, EAA, -x0, z0, spin_rate, SADiff), nrow = 1, ncol = 8)), 
-                                -0.0011801, 0.007989927)),
-        Pitch %in% c("Changeup", "Splitter", "Screwball", "Forkball") ~ 
-          ifelse(any(is.na(c(start_speed, IVB, -HB, EAA, -x0, z0, spin_rate, SADiff))), 
-                 NA_real_, 
-                 scale_TimStuff(predict(Off, matrix(c(start_speed, IVB, -HB, EAA, -x0, z0, spin_rate, SADiff), nrow = 1, ncol = 8)), 
-                                -0.002239657, 0.01043216)),
-        Pitch %in% c("Slider", "Slurve", "Sweeper", "Curveball", "Knuckle Curve", "Slow Curve", "Knuckleball") ~ 
-          ifelse(any(is.na(c(start_speed, IVB, -HB, EAA, -x0, z0, spin_rate, SADiff))), 
-                 NA_real_, 
-                 scale_TimStuff(predict(Break, matrix(c(start_speed, IVB, -HB, EAA, -x0, z0, spin_rate, SADiff), nrow = 1, ncol = 8)), 
-                                -0.004822031, 0.007912765)),
-        TRUE ~ NA_real_
-      )
-    } else {
-      case_when(
-        Pitch %in% c("Four-Seam Fastball", "Sinker", "Cutter", "Fastball") ~ 
-          ifelse(any(is.na(c(start_speed, IVB, HB, EAA, x0, z0, spin_rate, SADiff))), 
-                 NA_real_, 
-                 scale_TimStuff(predict(FB, matrix(c(start_speed, IVB, HB, EAA, x0, z0, spin_rate, SADiff), nrow = 1, ncol = 8)), 
-                                -0.0011801, 0.007989927)),
-        Pitch %in% c("Changeup", "Splitter", "Screwball", "Forkball") ~ 
-          ifelse(any(is.na(c(start_speed, IVB, HB, EAA, x0, z0, spin_rate, SADiff))), 
-                 NA_real_, 
-                 scale_TimStuff(predict(Off, matrix(c(start_speed, IVB, HB, EAA, x0, z0, spin_rate, SADiff), nrow = 1, ncol = 8)), 
-                                -0.002239657, 0.01043216)),
-        Pitch %in% c("Slider", "Slurve", "Sweeper", "Curveball", "Knuckle Curve", "Slow Curve", "Knuckleball") ~ 
-          ifelse(any(is.na(c(start_speed, IVB, HB, EAA, x0, z0, spin_rate, SADiff))), 
-                 NA_real_, 
-                 scale_TimStuff(predict(Break, matrix(c(start_speed, IVB, HB, EAA, x0, z0, spin_rate, SADiff), nrow = 1, ncol = 8)), 
-                                -0.004822031, 0.007912765)),
-        TRUE ~ NA_real_
-      )
-    }) %>%
     group_by(Pitch) %>%
     summarize(
       Pitches = n(),
@@ -237,7 +269,7 @@ summary_table <- function(game) {
       'VAA' = round(mean(VAA, na.rm = TRUE), digits = 1),
       'CSW%' = round(sum(team_fielding_id, na.rm = TRUE) / sum(!is.na(team_fielding_id)) * 100, digits = 1),
       'Whiff%' = round(sum(is_strike_swinging, na.rm = TRUE) / sum(swing, na.rm = TRUE) * 100, digits = 1),
-      'TimStuff' = round(mean(TimStuff, na.rm = TRUE), digits = 0)
+      'TimStuff+' = round(mean(TimStuff, na.rm = TRUE), digits = 0)
     ) %>%
     arrange(-Pitches)