Height_Age_and_Performance

#See results here: https://www.overleaf.com/read/qdybynqzdgns

//------------------------HEIGHT AND PERFORMANCE------------------------------------------------------------------------

# To build the indicator of performance we start by compute the number of victory and defeat of each player
number_win_by_name<-as.data.frame(table(atp_match$winner_name))
number_lose_by_name<-as.data.frame(table(atp_match$loser_name))

#we put this data in the same data frame
merge(number_win_by_name, number_lose_by_name, by ="Var1")

# we keep only  the line for which the payer has played at least 20 matches
win_lose_sup_by_name <- subset(win_lose_by_name, win_lose_by_name$Freq.x + win_lose_by_name$Freq.y > 20 )

#Then we compute the Ratio of performance win matches/ total nbr of matches
win_lose_sup_by_name$Ratio_w_l=win_lose_sup_by_name$Freq.x/(win_lose_sup_by_name$Freq.y+win_lose_sup_by_name$Freq.x)

# we draw the bowplot of the performance to observe its distribution among the player
boxplot(win_lose_sup_by_name$Ratio_w_l)

# we add the height of the player to our data frame and we keep only the distincts lines
size_and_name<-as.data.frame(atp_match[,c(11,13)])
names(size_and_name)[names(size_and_name) == 'winner_name'] <- 'Var1'
win_lose_size_name<-merge(win_lose_by_name, size_and_name, by ="Var1")
win_lose_size_name<-distinct(win_lose_size_name)

# we draw the qq plot to observe the correlation between the ratio of performance and the height + line of linear regression
ggplot(data = win_lose_size_name, aes(x = win_lose_size_name$Ratio_w_l, y = win_lose_size_name$winner_ht)) 
+ geom_point(color='red') 
+ geom_smooth(method = "lm", se = FALSE)
+ labs(x = "height (cm) ",
+ y = "Performance",
+ title = "Correlation between performance and size")

# we also add the Spearman test of correlation
res<-cor.test(win_lose_size_name$Ratio_w_l,win_lose_size_name$winner_ht, method="spearman")
res



//------------------------AGE AND PERFORMANCE------------------------------------------------------------------------

# we add the ageof the player to our data frame and we keep only the distincts lines
age_and_name<-as.data.frame(atp_match[,c(11,15)])
names(age_and_name)[names(age_and_name) == 'winner_name'] <- 'Var1'
win_lose_sup_by_name<-merge(win_lose_sup_by_name, age_and_name, by ="Var1")
win_lose_sup_by_name<-distinct(win_lose_sup_by_name)
View(win_lose_sup_by_name)

# we draw the qq plot to observe the correlation between the ratio of performance and the age+ line of linear regression
ggplot(data = win_lose_sup_by_name, aes(x = win_lose_sup_by_name$Ratio_w_l, y = win_lose_sup_by_name$winner_age))
+ geom_point(color='red') + geom_smooth(method = "lm", se = FALSE)+
labs(x = "Performance ",y = "age", title = "Correlation between performance and age")

# we also add the Spearman test of correlation
res<-cor.test(win_lose_sup_by_name$Ratio_w_l,win_lose_sup_by_name$winner_age, method="spearman")
res