Adding percentage labels to ggplot when using stat_count

Question

For some reason, I can't seem to be able to add correct proportion labels to a ggplot by using stat_count. The code below returns labels that display 100% for all categories, even though I'm using ..prop... Should I use something else instead of stat_count?

library(tidyverse)
diamonds %>% 
  ggplot(aes(color, fill=cut)) +
  geom_bar(position = 'fill') +
  stat_count(aes(label= scales::percent(..prop..)), 
             geom = 'text', position = position_fill(vjust = 0.5))

I know this can also be accomplished by calculating the percentage before feeding the data to ggplot (like below) but I have quite a bit of code which is using geom_bar and I would need to change all of it if I were to do it this way.

diamonds %>% 
  count(color, cut) %>% 
  group_by(color) %>% 
  mutate(pct=n/sum(n)) %>% 
  ggplot(aes(color, pct, fill=cut)) +
  geom_col(position = 'fill') +
  geom_text(aes(label=scales::percent(pct)), position = position_fill(vjust=0.5))

I just installed the latest ggplot2 version but I still have the same issue — Vlad C.
– Vlad C., Commented Aug 30, 2018 at 8:14

Z.Lin · Accepted Answer · 2018-08-30 07:55:52Z

10

You can do the calculations within geom_label(), if you don't want to change the geom_bar part:

diamonds %>% 
  ggplot(aes(color, fill=cut)) +
  geom_bar(position = 'fill') +
  geom_text(data = . %>% 
              group_by(color, cut) %>%
              tally() %>%
              mutate(p = n / sum(n)) %>%
              ungroup(),
            aes(y = p, label = scales::percent(p)),
            position = position_stack(vjust = 0.5),
            show.legend = FALSE)

edited Aug 30, 2018 at 7:55

answered Aug 30, 2018 at 7:36

Z.Lin

29.3k7 gold badges61 silver badges100 bronze badges

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

3 Comments

Vlad C. Over a year ago

Thank you, Z.Lin! Your solution works! Do you know why the default ..prop.. doesn't work (or what situations it was meant for)?

Z.Lin Over a year ago

Based on my understanding, ..prop.. gives the groupwise proportion. By default stat_count considers each x-axis position separately, and fill = cut means each "cut" value (Fair / Good / etc.) is considered a distinct group. Thus each color-cut combination becomes its own group, with proportion = 100%.

Luis Over a year ago

Hey, great answer! Simple and understandable!

Collectives™ on Stack Overflow

Adding percentage labels to ggplot when using stat_count

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