NAs Analysis

PreparingData.R

@@ -56,6 +56,26 @@ dim(laLigaPlayers.1920)
 laLigaPlayers.1920 <- aceptaTildes(laLigaPlayers.1920)
 
 allLaLigaPlayer <- list(laLigaPlayers.1718, laLigaPlayers.1819, laLigaPlayers.1920)
+
+####################
+# VALORES AUSENTES #
+####################
+na.var <- function(variable){
+  sum(is.na(variable) + 0)
+}
+nas <- apply(laLigaPlayers.1718, 2, na.var)
+prop.nas <- round(nas/dim(laLigaPlayers.1718)[1], 2)
+dataframe.na.1718 <- data.frame("No NAs" =  dim(laLigaPlayers.1718)[1] - nas,
+                                "NAs" = nas, "Porcentaje NAs" = 100*prop.nas)
+
+nas <- apply(laLigaPlayers.1819, 2, na.var)
+porc.nas <- round(nas/dim(laLigaPlayers.1819)[1], 2)
+dataframe.na.1819 <- data.frame("No NA" =  dim(laLigaPlayers.1819)[1] - nas,"NAs" = nas, "Porcentaje NAs" = 100*prop.nas)
+
+nas <- apply(laLigaPlayers.1920, 2, na.var)
+prop.nas <- round(nas/dim(laLigaPlayers.1920)[1], 2)
+dataframe.na.1920 <- data.frame("No NA" =  dim(laLigaPlayers.1920)[1] - nas,"NAs" = nas, "Porcentaje NAs" = 100*prop.nas)
+
 ##########################################################
 # Jugadores que CAMBIARON de club en una misma temporada #
 ##########################################################
@@ -177,6 +197,23 @@ rowsPIn2C
 # CONJUNTO de ENTRENAMIENTO # # CONJUNTO de TEST #
 ############################# ####################
 set.seed(5682)
-ind <- sample(2, nrow(laLigaPlayers), replace = FALSE, prob = c(0.8, 0.2))
+ind <- sample(2, nrow(laLigaPlayers), replace = TRUE, prob = c(0.8, 0.2))
 train <- laLigaPlayers[ind == 1, ]
 test <- laLigaPlayers[ind == 2, ] 
+
+############################################################################
+# Comparación de jugadores que juegan para 2 EQUIPOS en la MISMA temporada #
+############################################################################
+Nolito <- laLigaPlayers[rowsPIn2C[14,], ]
+variables <- paste(c("age.", "height.", "games.", "minutes.", "games_starts.",
+                     "xa.", "passes_pct.", "passes_total_distance.", "assisted_shots.", "passes_switches.", "pass_targets.", "passes_received", "passes_received_pct.", "passes_pressure.",
+                     "xg.", "npxg.",  "shots_on_target_pct.", "goals_per_shot.", "pens_made.", "pens_att.", 
+                     "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."), "1920", sep = "")
+Nolito <- Nolito %>% select(contains(variables))
+Nolito <- t(Nolito)
+# 177: Celta
+# 178: Sevilla
+Nolito <- data.frame(lapply(Nolito, as.numeric))
+