Harmonize width and location of horizontal bars across facets with different number of bars

[werkstattcodes]

First, many thanks again to everyone involved with this wonderful package.

I regularly make plots like the one below. While it conveys all the necessary info, its format is (imho) not really fit for publication.

What bothers me is that
a) bar widths are not the same (=should all have the width of the bars in the facet with most bars, i.e. europe),
b) that the bars in the different facets are not horizontally aligned (i.e. the bar BR in the facet south america should have the same vertical position as the bar ES in the facet europe). I would have hoped that facet_row would do the trick, but it seems not to be the case.

I might be mistaken, but it seems to me that the plot below is a fairly regularly required format. It would be great if there is a convenient way to produce it.

Many thanks!

library(tidyverse)
df <- data.frame(
  stringsAsFactors = FALSE,
         continent = c("europe","europe","europe",
                       "europe","asia","asia","asia","south america"),
           country = c("AT", "DE", "DK", "ES", "PK", "IN", "TW", "BR"),
             value = c(10L, 15L, 20L, 25L, 17L, 17L, 13L, 5L)
)

df %>% 
  ggplot()+
  geom_bar(aes(y=country,
               x=value),
           orientation = "y",
           stat="identity")+
  facet_wrap(vars(continent),
             scales = "free_y")

^{Created on 2020-08-16 by the reprex package (v0.3.0)}

[teunbrand]

This is not really a solution but a workaround for this problem if you know in advance what the maximum number of categories on the y axis is. It uses the ability of limits to be functions to encode dummy categories that can be filtered out at the break level.

library(tidyverse)
df <- data.frame(
  stringsAsFactors = FALSE,
  continent = c("europe","europe","europe",
                "europe","asia","asia","asia","south america"),
  country = c("AT", "DE", "DK", "ES", "PK", "IN", "TW", "BR"),
  value = c(10L, 15L, 20L, 25L, 17L, 17L, 13L, 5L)
)

maxnumcat <- 4

df %>% 
  ggplot()+
  geom_bar(aes(y=country,
               x=value),
           orientation = "y",
           stat="identity")+
  scale_y_discrete(
    limits = function(x) {
      y <- paste0("dummy", seq_len(maxnumcat))
      c(y[seq_len(maxnumcat - length(x))], x)
    },
    breaks = function(x) {
      x[!startsWith(x, "dummy")]
    }
  ) +
  facet_wrap(vars(continent),
             scales = "free_y")

^{Created on 2020-08-18 by the reprex package (v0.3.0)}

[werkstattcodes]

That's really a neat solution. Many thank!
Irrespective of your solution, I could still imagine that a 'ready-made' option provided by ggplot (or ggforce?) would be very helpful.

[thomasp85]

It is unlikely that we will add support for this anytime soon due to limited developer bandwidth, though I admit I understand your use case.

[werkstattcodes]

Thank you, Thomas.