ggplot, pie chart and different radii in facet_wrap

Question

I would like to create pie charts with different radii in separate facets. How can I modify the following code (which produces equal radii) so the variable total is th rasdius for each slice?

library(tidyverse)
mydf <- data_frame(value=c(1:3,1:3),
                   total=rep(c(10,20),times=3),
                           cond=rep(c("x","y","z"),times=2),
                   group=rep(c("a","b"),each=3))
mydf %>% ggplot(aes(x="",y=value,fill=cond)) +
                  geom_bar(stat="identity",width=1) +
  facet_wrap(~ group) +
  coord_polar("y")

Answer 1

This an old question but I wanted to do the same thing: The pie radius should vary with the group total. First of all, I use geom_bar(position = "fill") so that the two group bars stack up to the same height. Then I use the group totals as a width aesthetic: ggplot(aes(x = total/2, width = total)). Using half the total for the x aesthetic ensures that we get proper pies not doughnuts. So, the following code works for me although I'm not sure about how much of this is rather a hack than a proper solution (ggplot issues a warning if we move the width aesthetic into the geom_barcall: geom_bar(aes(width = total))).

library(tidyverse)

mydf <- tibble(group = rep(c("group a", "group b"), each = 3),
               cond = rep(c("x", "y", "z"), times = 2),
               value = c(1, 2, 3, 2, 4, 6)) %>% 
    group_by(group) %>% 
    add_tally(value, name = "total") %>% 
    ungroup() %>%
    mutate(label = sprintf("total = %d", total)) %>% 
    print()
#> # A tibble: 6 x 5
#>   group   cond  value total label     
#>   <chr>   <chr> <dbl> <dbl> <chr>     
#> 1 group a x         1     6 total = 6 
#> 2 group a y         2     6 total = 6 
#> 3 group a z         3     6 total = 6 
#> 4 group b x         2    12 total = 12
#> 5 group b y         4    12 total = 12
#> 6 group b z         6    12 total = 12

mydf %>% ggplot(aes(x = total/2, y = value, fill = cond, width = total)) +
    geom_bar(stat = "identity", position = "fill", color = "white") +
    facet_wrap(~ group + label, strip.position = "bottom") +
    coord_polar("y", start = 0, direction = -1) +
    theme_bw(base_size = 12) +
    theme(axis.title = element_blank(),
          axis.ticks = element_blank(),
          axis.text = element_blank(),
          panel.grid = element_blank(),
          panel.border = element_blank(),
          legend.title = element_text(size = 14), 
          strip.background = element_rect(fill = NA, colour = NA),
          strip.text = element_text(size = 16))

^{Created on 2020-02-13 by the reprex package (v0.3.0)}

To better understand how this works, we can remove coord_polar() and the theme() calls to have a look at the underlying bar chart:

mydf %>% ggplot(aes(x = total/2, y = value, fill = cond, width = total)) +
    geom_bar(stat = "identity", position = "fill", color = "white") +
    facet_wrap(~ group + label, strip.position = "bottom") +
    theme_bw(base_size = 12) 

Answer 2

Your question is a bit unclear. Something like that?

library(tidyverse)
mydf <- data_frame(value = c(1:3, 1:3),
                   total = rep(c(10, 20), times = 3),
                   cond = rep(c("x", "y", "z"),times = 2),
                   group = rep(c("a", "b"), each = 3))
mydf %>% ggplot(aes(x = "", y = value, fill = cond)) +
    geom_bar(stat = "identity", width = 1) +
    facet_wrap(total ~ cond) +
    coord_polar("x")