Using Dplyr "group_by" and "Summarise" and a Custom Function to Calculate the Mode of Several Groups

Question

Apparently dplyr's summarise function doesn't include an option for "mode". Based on the simple data frame example below, I would like to determine the mode, or most frequently repeating number, for each group of "Category." So for group "A", the mode is 22, for "B", it's 12 and 14, and there is no repeating number for "C".

I found some examples of functions online, but none addressed the situation when there are no repeating numbers in a group. Is there a need for a custom function, or is there a mode option somewhere? I don't want to rely on any other specialized packages just for their mode function. It would be nice to find an elegant and simple solutioin using a combination of base R, dplyr, tidy, etc.

If a custom function is used, it will have to work when there are no repeating numbers, as well as when there are more than one equally repeating number.

Any help would be greatly appreciated! This seems like it should be an easy solutioin in R, so I was surprised to learn that there is no simple summarise_each(funs(mode)... option.

If a custom function is used, please break it down with explanations. I'm still relatively new to R functions.

Category<-c("A","B","B","C","A","A","A","B","C","B","C","C")
Number<-c(22,12,12,8,22,22,18,14,10,14,1,3)
DF<-data.frame(Category,Number)

akrun · Accepted Answer · 2016-07-05 05:33:09Z

5

We can use

 Mode <- function(x) {
  ux <- unique(x)
  if(!anyDuplicated(x)){
      NA_character_ } else { 
     tbl <-   tabulate(match(x, ux))
     toString(ux[tbl==max(tbl)])
 }
}

DF %>%
   group_by(Category) %>%
   summarise(NumberMode = Mode(Number))
#  Category NumberMode
#    <fctr>      <chr>
#1        A         22
#2        B     12, 14
#3        C       <NA>

edited Jul 5, 2016 at 5:33

answered Jul 5, 2016 at 3:23

akrun

891k38 gold badges590 silver badges700 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Mike Over a year ago

Thanks for the fast reply. I tried this function before, but for group "C", which doesn't have any repeating numbers, it returns an 8. Isn't there a way to return something to indicate when there isn't a repeating number? Also, group "B" has two numbers that each repeat twice, so I imagine it returns the lowest number by default. Is there a way to return both numbers if there is more than one mode? This might be asking too much...but I would really like to know when there are no repeating numbers if possible...

akrun Over a year ago

@Mike What do you want for group C?

Mike Over a year ago

That's great, exactly what I wanted. Thanks!

Collectives™ on Stack Overflow

Using Dplyr "group_by" and "Summarise" and a Custom Function to Calculate the Mode of Several Groups

1 Answer 1

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related