Individual binwidths in faceted histogram on ggplot2

Question

I do a series of histograms with facet_grid and I want every histogram in the grid to have the same number of classes, in the example below e.g. 6 classes. The problem in this example below is that binwidth = diff(range(x$data))/6) defines the classes according to the overall range of a, b and c, i.e. defines one binwidth for all three facets.

How do I define binwidth individually for the facets a, b and c?

require("ggplot2")

a <- c(1.21,1.57,1.21,0.29,0.36,0.29,0.93,0.26,0.28,0.48,
       0.12,0.38,0.83,0.82,0.41,0.69,0.25,0.98,0.52,0.11)
b <- c(0.42,0.65,0.17,0.38,0.44,0.01,0.01,0.03,0.15,0.01)
c <- c(1.09,3.55,1.07,4.55,0.55,0.11,0.72,0.66,1.22,3.04,
       2.01,0.64,0.47,1.33,3.44)

x <- data.frame(data = c(a,b,c), variable = c(rep("a",20),rep("b",10),rep("c",15)),area="random")

qplot(data, data = x, geom = "histogram", binwidth = diff(range(x$data))/6) +
  facet_grid(area~variable, scales = "free")

Answer 1

This is not optimal but you can do the histogram in different layers:

ggplot(x, aes(x=data)) +
   geom_histogram(data=subset(x, variable=="a"), binwidth=.1) +
   geom_histogram(data=subset(x, variable=="b"), binwidth=.2) +
   geom_histogram(data=subset(x, variable=="c"), binwidth=.5) +
   facet_grid(area~variable, scales="free")

Answer 2

One way is to pre-summarize your data in the way you want it, then to create the plot.

In your case, you need to bin your variables using the function cut(). The package dplyr is convenient for this, because it allows you to specify a mutate function for each group of your data:

library(dplyr)

zz <- x %>%
  group_by(variable) %>%
  mutate(
    bins = cut(data, breaks=6)
  )

qplot(bins, data = zz, geom = "histogram", fill=I("blue")) +
  facet_grid(area~variable, scales = "free") +
  theme(axis.text.x = element_text(angle=90))