How do I create multiple boxplots whilst also making them proportional using geom_boxplot within ggplot2?

Question

I understand that varwidth = TRUE within ggplot2's geom_boxplot can be used to create proportional boxplots so that each boxplot also summarises the number of points within a plot. However, I am struggling to maintain proportional boxplot sizes whilst also producing multiple plots?

Using the diamonds df in ggplot2 I am trying to reproduce the below image.

What I have tried is this:

ggplot(data = diamonds, mapping = aes(x = carat, y = price)) + geom_boxplot(mapping = aes(group = cut_number(carat, 20), varwidth = TRUE))

I tried to use cut_number to achieve the multiple, and var-width for proportionality. I'd also like to have the boxplots show vertically. I have been stumped on this for hours and have looked online but to no avail. Any tips?

Answer 1

Since ggplot2 v3.3.0, the orientation of layers are autodetected. In some cases, this detection can be incorrect. You can force the orientation by adding it as an argument to a layer:

library(ggplot2)

ggplot(data = diamonds, mapping = aes(x = carat, y = price)) + 
  geom_boxplot(mapping = aes(group = cut_number(carat, 20)),
               orientation = "x")

^{Created on 2020-05-01 by the reprex package (v0.3.0)}