Partially color histogram in R

Question

I have a histogram as shown in the picture. I want the bars in the two regions to be coloured red and green respectively, i.e., the bars from 0 to the first black border on the left should be coloured red and the bars in the second region should be coloured green. Can this be done in R? The code used to get the histogram is

hist(pr4$x[pr4$x[,1]>-2,1],breaks=100)

Answer 1

The best way to do this is to allow hist to do the calculations for you but then use hist (again) to do the actual plotting. Here's an example:

set.seed(1)
x <- rnorm(1000)
h <- hist(rnorm(1000), breaks=50, plot=FALSE)
cuts <- cut(h$breaks, c(-Inf,-.5,1.75,Inf))
plot(h, col=cuts)

The .5 and 1.75 in the last line are the threshold values where you want to have different colors.

Note: my original answer used barplot, but that strategy makes the axes kind of difficult to work with.

Here's the result:

Answer 2

Here's the method I mentioned in comments:

Make some test data (you should do this in your question!)

test = runif(10000,-2,0)

get R to compute the histogram but not plot it:

h = hist(test, breaks=100,plot=FALSE)

Your histogram is divided into three parts:

ccat = cut(h$breaks, c(-Inf, -0.6, -0.4, Inf))

plot with this palette, implicit conversion of factor to number indexes the palette:

plot(h, col=c("white","green","red")[ccat])

ggplot2 solutions are also available.

Answer 3

Try this:

hist(pr4$x[pr4$x[,1]>-2,1],breaks=100, col = c(rep("white", 69), rep("green", 15), rep("red", 16)))

You might have to adjust the repeat number for each color depending on the number of breaks within the histogram.