Can you stack boxplots in ggplot2?

Question

This is an example of my plot. Notice how the boxplots are next to each other. I want to figure out if there is a way to stack the box plots on top of each other.

Answer 1

Update Because it was a challenging task I got a little further: The main hint was in the comments of Gregor Thomas. I searched a little and found this geom_boxplot: map alpha levels to whiskers & outliers

The result: Plot A can be shown as plot B with the code below:

library(ggplot2)
library(cowplot)

df <- diamonds

# defining individual alpha levels
# learned here https://stackoverflow.com/questions/34618517/geom-boxplot-map-alpha-levels-to-whiskers-outliers
numVals = length(unique(iris$Species))
avals = seq(0.1, 0.4, length.out=numVals)

avalsHex = paste0("#000000", toupper(as.hexmode(round(avals*255))))

# the plots with some data tweaking
p1 <- df %>% 
  mutate(color = fct_lump(color, 2),
         mutate = fct_reorder(color, depth)) %>% 
  ggplot(aes(cut, depth))+
  geom_boxplot(aes(fill=color, alpha=color)) +
  scale_alpha_manual(values = avals) +
  scale_colour_manual(values = avalsHex) +
  theme_void()

p2 <- df %>% 
  mutate(color = fct_lump(color, 2),
         mutate = fct_reorder(color, depth)) %>% 
  ggplot(aes(cut, depth))+
  geom_boxplot(aes(fill=color, alpha=color), position = "identity") +
  scale_alpha_manual(values = avals) +
  scale_colour_manual(values = avalsHex) +
  theme_void()

plot_grid(p1, p2, labels = "AUTO")

First answer: I am not sure, but you can achieve a kind of stacking with coord_flip. See my example with diamonds dataset. This means to just show the boxplots without the dots.

ggplot(diamonds, aes(x = cut, y = depth, fill = color)) +
  geom_boxplot() +
  coord_flip()