I have customer call data and I want to get the median and mode for the call success rate for different subgroups.

My variables are: Customer ID, Employment Status (Retired, Employed, Unemployed), count of calls, count of successful calls, call success rate (successful calls/number of calls)

enter image description here

So far I worked in Excel to get the average call reachability. Mode and median are not easy to calculate with Excel, because it can only do it for up to 255 recods I believe. I know that the summary command in R provides you with summary statistics for each variable but what if you want to get the median and mode of each subgroup. Meaning: Median and Mode of call succcess rate for Retired, Employed and Unemployed customers separately? Why I want to calculate the Median and Mode, although I have the average already? Because both are more resistant to outliers. But I am open for discussion if you think it is not necessary :)


assuming your data is stored in an object named df you can do:

tapply(df$S_Calls, df$Emp_Stat, median)

As for the mode, oddly enough R does not have a built in function for that. You could define one yourself using:

mode_stat <- function(x) {
  ux <- na.omit(unique(x))
  ux[which.max(tabulate(match(x, ux)))]

and then do in a similar fashion:

tapply(df$S_Calls, df$Emp_Stat, mode_stat)

Since you mentioned that you have the data in excel, first import it in R:

df <- readxl::excel_sheets(path = "file.xlsx")

then calculate median per Emp_Stat:

sapply(unique(df$Emp_Stat), function(x){
    sapply(df[df$Emp_Stat == x, ], median)

and for mode:

# for every Emp_Stat
sapply(unique(df$Emp_Stat), function(x1){
    # for every column
    sapply(df[df$Emp_Stat == x1, ], function(x2){

Ps. What I mentioned and the answer Iyar Lin provided are essentially the same, just using different functions.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.