Combine statistics of the 3 seasons

PreparingData.R

@@ -27,11 +27,22 @@ aceptaTildes <- function(datos) {
   return(datos)
 }
 
+decimalplaces <- function(x) {
+  ifelse(abs(x - round(x)) > .Machine$double.eps^0.5,
+         nchar(sub('^\\d+\\.', '', sub('0+$', '', as.character(x)))),
+         0)
+}
+
 #############
 # LIBRERIAS #
 #############
 library(dplyr) # to inner join 
+library(tidyverse) # filter(), select(), contains()
 
+############
+# GLOBALES #
+############
+fusion <- 0
 
 ######################################
 #          LECTURA de DATOS          #
@@ -102,6 +113,7 @@ rowsPIn2CGlobal
 ####################################################
 # Fusion de las muestras de jugadores en 2 equipos #
 ####################################################
+if(fusion == 1) {
   pos <- which(rowsPIn2CGlobal[,6] == "17-18")
   r1 <- as.numeric(rowsPIn2CGlobal[pos, 2])
   r2 <- as.numeric(rowsPIn2CGlobal[pos, 3])
@@ -170,6 +182,7 @@ for (name in names) {
     laLigaPlayers.1920 <- laLigaPlayers.1920[-auxPos,]
   }
   laLigaPlayers.1920 <- rbind(laLigaPlayers.1920, players)
+}
 
 ##############
 # INNER JOIN #
@@ -256,3 +269,37 @@ Youssef <- t(Youssef)
 jugador <- laLigaPlayers[rowsPIn2C[9,], ]
 jugador <- jugador %>% select(contains(variables))
 jugador <- t(jugador)
+
+###################################################################
+# Combinacion de estadisticas por temporada en una unica variable #
+###################################################################
+# Seleccionamos unicamente las variables del estudio
+# No OLVIDAR "age.", "height.",
+variables <- c("xa.", "passes_completed.", "passes.", "passes_pct.", "passes_total_distance.", "assisted_shots.", "passes_switches.", 
+               "pass_targets.", "passes_received.", "passes_received_pct.", "passes_pressure.",
+               "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.", 
+               "goals_against_per90_gk.", "pens_saved.", "pens_allowed.", "shots_on_target_against.", "saves.", "save_pct.")
+globalColNames <- c()
+for (variable in variables) {
+  variable <- gsub("\\.", "", variable)
+  globalColNames <- c(globalColNames, variable)
+}
+variablesWithSeason <- c(paste(variables, "1718", sep = ""), paste(variables, "1819", sep = ""), paste(variables, "1920", sep = ""))
+
+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) {
+    aux2 <- round(rowSums(aux)/3, decimalplaces(aux[1,1]))
+  }
+  trainStudy <- cbind(trainStudy, aux2)
+}
+trainStudy <- as_tibble(trainStudy)
+names(trainStudy) <- globalColNames