data<-read.csv("Why are employees leaving.csv", header = TRUE)
#left job (how manypeople ??)
sum(data$left)
#creation of pi chart 
pie(c(sum(data$left),nrow(data)-sum(data$left)), labels = unique(data$left), col = rainbow(3))

# analysis division wise data
left_div<-aggregate(. ~ data$Division, data=data, sum)
#analysis division wise how many people left 
left_div1<-aggregate(data$left ~ data$Division, data= data, sum)

# Display Division wise average data 

divisional_data<-aggregate(data, by=list(data$Division), mean)

# plot data between satisfaction and latest evauation 
plot(data$satisfaction_level, data$last_evaluation)

#Analyzies person who left and why?

left_data<-data[which(data$left==1), names(data)]
plot(left_data$satisfaction_level, left_data$last_evaluation)
  
# Assigned individual cluster name 

category<-c(rep("NA",nrow(left_data)-1))
for (i in 1:nrow(left_data)){ 
  if(left_data$satisfaction_level[i] >0.7 & left_data$last_evaluation[i] > 0.7) {
    category[i]<- "winner"
  }
  else if(left_data$satisfaction_level[i] <0.2 & left_data$last_evaluation[i] > 0.7){
    category[i]<- "unfulfilled"
  }
  else if(left_data$satisfaction_level[i] <0.5 & left_data$last_evaluation[i] < 0.6){
    category[i]<- "badmatch"
  }
}
left_data<-cbind(left_data,category)
sum(left_data$category=="winner")
sum(left_data$category=="badmatch")
sum(left_data$category=="unfulfilled")
View(left_data)

#Division wise analysis for winner category 

t1<-aggregate(left_data$left[which(left_data$category=="winner")], by= list(left_data$Division[which(left_data$category=="winner")]), FUN = sum)