CODEWITHSUNDEEP
rdplyr
Jim
Jim

Reputation: 725

R group_by return number of largest unique type

Suppose I have this data set:

df <-  data.frame(c('a', 'a', 'a', 'a', 'a', 'a', 'b', 'b', 'b', 'b', 'b', 'b', 'b'),
  c('c', 'c', 'd', 'e', 'f', 'c', 'e', 'f', 'f', 'f', 'g', 'h', 'f')
    ) %>% setNames(c('type', 'value'))

   type value
1     a     c
2     a     c
3     a     d
4     a     e
5     a     f
6     a     c
7     b     e
8     b     f
9     b     f
10    b     f
11    b     g
12    b     h
13    b     f

I'd like to perform some kind of command as follows:

df %>% group_by(type) %>% 
  summarise_all(funs(largest_group_size))

This would ideally produce a table with the largest number of any value for a and b.

  type largest_group_size
1    a                  3
2    b                  4

This table would have:

Ideally, I'd like to go a step further and calculate the percentage that the largest group is of the whole by type. So (largest_group_size / n()).

Upvotes: 0

Views: 47

Answers (1)

CMcMaster
CMcMaster

Reputation: 21

In two group_by steps:

df %>% 
  group_by(type, value) %>% 
  summarise(groups = n()) %>% 
  group_by(type) %>% 
  summarise(largest_group = max(groups),
            as_percentage = largest_group / sum(groups))

This gives:

  type  largest_group as_percentage
  <fct>         <dbl>         <dbl>
1 a                 3         0.5  
2 b                 4         0.571

There is probably a more efficient way, but this is how I would do this in a hurry.

Upvotes: 2

Related Questions