Distribution of positions of the 3 seasons

PreparingData.R

@@ -33,6 +33,49 @@ decimalplaces <- function(x) {
          0)
 }
 
+# Returns the count of a variable from 3 seasons for each position with a bar plot.
+plotFor3SeasonsForPos <- function(data, variable) {
+  par(mfrow = c(1, 3))
+  dfs <- list()
+  maxPInPos <- 0
+  for (i in 1:3) {
+    season <- switch(i, ".1718", ".1819", ".1920")
+    var <- select(data, contains(paste(variable, season, sep = "")))
+    GK <- var %>% filter(str_detect(var[,1], 'GK'))
+    numGK <- dim(GK)[1]
+    DF <- var %>% filter(str_detect(var[,1], 'DF'))
+    numDF <- dim(DF)[1]
+    MF <- var %>% filter(str_detect(var[,1], 'MF'))
+    numMF <- dim(MF)[1]
+    FW <- var %>% filter(str_detect(var[,1], 'FW'))
+    numFW <- dim(FW)[1]
+    
+    df <- data.frame(
+      position = c("Porteros", "Defensas", "Mediocentros", "Delanteros"),
+      numberOf = c(numGK, numDF, numMF, numFW)
+    )
+    dfs[[i]] <- df
+    if(maxPInPos < max(df$numberOf)) {
+      maxPInPos <- max(df$numberOf)
+    }
+  }
+  
+  for (i in 1:3) {
+    season <- switch(i, ".1718", ".1819", ".1920")
+    bp <- barplot(height = dfs[[i]]$numberOf, names = dfs[[i]]$position, 
+                  #col = rgb(0.8,0.1,0.1,0.6),
+                  col = "#9BCD9B",
+                  xlab = "Posiciones", 
+                  ylab = "Cantidad", 
+                  main= paste(c("Distribución posiciones ", season, "(Validación)"), sep = ""),
+                  ylim = c(0, maxPInPos + 5)
+    )
+    prop <- round((dfs[[i]]$numberOf/sum(dfs[[i]]$numberOf))*100, 2)
+    cat(prop, "\n")
+    tabla <- table(rep(c("GK", "DF", "MF", "FW"), times = dfs[[i]]$numberOf))
+  }
+}
+
 #############
 # LIBRERIAS #
 #############
@@ -205,6 +248,23 @@ for (name in playersIn2Clubs) {
 rownames(rowsPIn2C) <- playersIn2Clubs
 rowsPIn2C
 
+######################
+# VARIABLES CREATION #
+######################
+
+# AERIALS
+aerials <- laLigaPlayers %>% select(all_of(c("aerials_won.1718", "aerials_lost.1718")))
+aerials_contested <- rowSums(aerials)
+laLigaPlayers <- cbind(laLigaPlayers, "aerials_contested.1718" = aerials_contested)
+
+aerials <- laLigaPlayers %>% select(all_of(c("aerials_won.1819", "aerials_lost.1819")))
+aerials_contested <- rowSums(aerials)
+laLigaPlayers <- cbind(laLigaPlayers, "aerials_contested.1819" = aerials_contested)
+
+aerials <- laLigaPlayers %>% select(all_of(c("aerials_won.1920", "aerials_lost.1920")))
+aerials_contested <- rowSums(aerials)
+laLigaPlayers <- cbind(laLigaPlayers, "aerials_contested.1920" = aerials_contested)
+
 ##############################################
 # Limpieza muestras erroneas temporada 18-19 #
 ##############################################
@@ -228,14 +288,15 @@ playersR.1819 <- playersR.1819 %>% select(all_of(c("player", "squad.1819", varia
 
 # Ibai Gomez, Munir El Haddadi, Ruben Sobrino, Jeison Murillo, Facundo Roncaglia, Takashi Inui, Ruben Vezo
 # 4, 8, 13, 66, 72, 84, 109
+jugadoresRepes <- laLigaPlayers[rowsPIn2C[-c(1:4, 6:8),], ]
+laLigaPlayerSplit <- laLigaPlayers[-rowsPIn2C[-c(1:4, 6:8),], ]
 laLigaPlayers <- laLigaPlayers[-c(4, 8, 13, 66, 72, 109), ]
+
 ############################# ####################
 # CONJUNTO de ENTRENAMIENTO # # CONJUNTO de TEST #
 ############################# ####################
 set.seed(5682)
 library(rsample)
-jugadoresRepes <- laLigaPlayers[rowsPIn2C[-c(1:4, 6:8),], ]
-laLigaPlayerSplit <- laLigaPlayers[-rowsPIn2C[-c(1:4, 6:8),], ]
 car_split <- initial_split(laLigaPlayerSplit, prop = 8/10, strata = "position.1920")
 train <- training(car_split)
 train <- rbind(train, jugadoresRepes)
@@ -244,6 +305,8 @@ test <- testing(car_split)
 plotFor3SeasonsForPos(laLigaPlayers, "position")
 plotFor3SeasonsForPos(train, "position")
 plotFor3SeasonsForPos(test, "position")
+# Tamaño: 1024x768
+
 ############################################################################
 # Comparación de jugadores que juegan para 2 EQUIPOS en la MISMA temporada #
 ############################################################################
@@ -280,7 +343,7 @@ variables <- c("xa.", "passes_completed.", "passes.", "passes_pct.", "passes_tot
                "xg.", "npxg.",  "shots_total.", "shots_on_target.", "shots_on_target_pct.", "goals_per_shot.", "pens_made.", "pens_att.", 
                "dribbles.", "dribbles_completed.", "dribbles_completed_pct.", "goals_assists_per90.",
                "sca_passes_dead.", "gca_passes_dead.", "sca_passes_live.", "gca_passes_live.", "sca_dribbles.", "gca_dribbles.", "sca_fouled.", "gca_fouled.",
-               "passes_intercepted.", "ball_recoveries.", "pressure_regains.", "fouls.", "tackles_won.", "tackles.", "aerials_won.", "aerials_lost.", "aerials_won_pct.", 
+               "passes_intercepted.", "ball_recoveries.", "pressure_regains.", "fouls.", "tackles_won.", "tackles.", "aerials_won.", "aerials_lost.", "aerials_contested.", "aerials_won_pct.", 
                "goals_against_per90_gk.", "pens_saved.", "pens_allowed.", "shots_on_target_against.", "saves.", "save_pct.")
 globalColNames <- c()
 for (variable in variables) {
@@ -293,7 +356,6 @@ trainStudyAux <- train %>% select(all_of(c("player", variablesWithSeason)))
 
 trainStudy <- NULL
 for (variable in variables) {
-  cat(variable)
   aux <- trainStudyAux %>% select(contains(variable))
   aux <- data.frame(lapply(aux, as.numeric))
   if (dim(aux)[2] == 3) {
@@ -301,5 +363,9 @@ for (variable in variables) {
   }
   trainStudy <- cbind(trainStudy, aux2)
 }
+trainStudy <- cbind(train %>% select(all_of(c("player"))), trainStudy)
 trainStudy <- as_tibble(trainStudy)
-names(trainStudy) <- globalColNames
+names(trainStudy) <- c("player", globalColNames)
+
+
+  
\ No newline at end of file