Scatter plots with correlation coefficient

BivariateAnalysis.R

@@ -32,7 +32,9 @@ library(correlation) # Correlation matrix table
 library(RColorBrewer)
 library(yarrr)
 library(ggrepel)
-library(PerformanceAnalytics)
+library(PerformanceAnalytics) # o
+library(ggpubr)
+
 
 # Seleccion de las varibles de ESTUDIO
 # variables <- paste(c("age.", "height.", "goals_assists_per90.", "xg.", "npxg.", "xa.", "shots_on_target_pct.", "passes_pct.", 
@@ -50,7 +52,7 @@ variables <- paste(c("age.", "height.",
                      "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."), "1718", sep = "")
+                     "goals_against_per90_gk.", "pens_saved.", "pens_allowed.", "shots_on_target_against.", "saves.", "save_pct."), "1819", sep = "")
 
 trainStudy <- train %>% select(starts_with(variables))
 trainStudy <- data.frame(lapply(trainStudy, as.numeric))
@@ -71,7 +73,46 @@ correlation::correlation(trainStudyStand,
 # Porcentaje SI/NO
 passes <- trainStudy[, 4:6]
 chart.Correlation(passes, histogram = TRUE, pch = 18)
+scatterForCorr <- function(datos, variableX, variableY, season) {
+  X <- paste(variableX, ".", season, sep = "")
+  Y <- paste(variableY, ".", season, sep = "")
+  medianX <- median(eval(parse(text = paste("datos$",X, sep = ""))))
+  maxY <- min(eval(parse(text = paste("datos$",Y, sep = ""))))
+  
+  sp <- ggscatter(datos, x = X, y = Y,
+                  add = "reg.line",  # Add regressin line
+                  add.params = list(color = "#B23AEE", fill = "lightgray"), # Customize reg. line
+                  conf.int = TRUE # Add confidence interval
+  )
+  sp + stat_cor(method = "pearson", label.x = medianX, label.y = maxY)
+}
+# 500x406
+scatterForCorr(trainStudy, "passes","passes_completed", "1819")
+scatterForCorr(trainStudy, "passes_completed", "passes_pct", "1819")
+scatterForCorr(trainStudy, "passes", "passes_pct", "1819")
+
+scatterForCorr(trainStudy, "pass_targets", "passes_received", "1819")
+scatterForCorr(trainStudyStand, "passes_received", "passes_received_pct", "1819")
+scatterForCorr(trainStudyStand, "pass_targets", "passes_received_pct", "1819")
+
+scatterForCorr(trainStudy, "shots_total", "shots_on_target", "1819")
+scatterForCorr(trainStudyStand, "shots_on_target", "shots_on_target_pct", "1819")
+scatterForCorr(trainStudyStand, "shots_total", "shots_on_target_pct", "1819")
+
+scatterForCorr(trainStudy, "aerials_lost", "aerials_won", "1819")
+scatterForCorr(trainStudyStand, "aerials_won", "aerials_won_pct", "1819")
+scatterForCorr(trainStudyStand, "aerials_lost", "aerials_won_pct", "1819")
+
+scatterForCorr(trainStudy, "dribbles", "dribbles_completed", "1819")
+scatterForCorr(trainStudyStand, "dribbles_completed", "dribbles_completed_pct", "1819")
+scatterForCorr(trainStudyStand, "dribbles", "dribbles_completed_pct", "1819")
+
+scatterForCorr(trainStudy, "shots_on_target_against", "saves", "1819")
+scatterForCorr(trainStudyStand, "saves", "save_pct", "1819")
+scatterForCorr(trainStudyStand, "shots_on_target_against", "save_pct", "1819")
 
+# Add correlation coefficient
+sp + stat_cor(method = "pearson", label.x = , label.y = 30)
 #######################
 # ANALISIS BIVARIANTE #
 #######################