Set binwidth to a factor of the x variable range in ggplot geom_histogram

Question

I'd like to create histograms of multiple variables, but for each facet I want the binwidth to be the size of the range between min and max for the x variable, divided by 20.

Using iris as an example this is what I have tried

irislong= pivot_longer(iris, cols = -Species)

ggplot(irislong, aes(x= value))+
  geom_histogram(binwidth = (max("value")- min("value"))/20, colour= "black") # also tried without "" around value
  facet_wrap(~ name, scales = "free")

Answer 1

You could pre-calculate each binwidth per group and add them in the binwidth aesthetics like this:

library(tidyr)
irislong= pivot_longer(iris, cols = -Species)

library(ggplot2)
library(dplyr)
p <- irislong %>%
  group_by(name) %>%
  mutate(width = (max(value) - min(value))/20) %>%
  ungroup() %>%
  ggplot(aes(x = value, binwidth = width, group = name)) +
  geom_histogram(colour = 'black') +
  facet_wrap(~name, scales = 'free')

p

# Check binwidths in layer
q <- ggplot_build(p)

q$data[[1]] %>%
  group_by(group) %>%
  summarise(binwidth = unique(binwidth))
#> # A tibble: 4 × 2
#>   group binwidth
#>   <int>    <dbl>
#> 1     1    0.295
#> 2     2    0.12 
#> 3     3    0.18 
#> 4     4    0.12

^{Created on 2022-11-23 with reprex v2.0.2}