Side-by-side and vertically stacked boxplots in R

Question

I have 3 different populations (staff types 1, 2, and 3) that are being evaluated along 3 variables (degree, weighted.degree, eccentricity) and am looking to view these in a single set of boxplots.

These boxplots should be placed side-by-side for each staff type for a given variable, and with 3 vertical stacked graphs by variable: something like a 3 x 3 matrix of boxplots.

My current code is producing the desired vertical stacking for the variables, but within each vertically stacked graph I have 1 aggregate boxplot where there should be 3 side-by-side (1 for each of the 3 staff types).

qp <- ggplot(data, aes(stafftype., value, fill=stafftype.))
+ geom_boxplot() + facet_grid(variable~., scales = "free_y")

Also note that each of the 3 side-be-side boxplots should be colored by stafftype, which I've been unsuccessful with.

resulting boxplot graph

Answer 1

Here's an example using the built-in iris data frame. We remove one of the measure columns so that we'll have three instead of four and then melt the data frame so that the three measure columns are stacked into "long" format. Species is analogous to stafftype and variable is analogous to the three evaluation variables (degree, weighted.degree, eccentricity).

library(reshape2)

ggplot(melt(iris[,-1], id.var="Species")) +
  geom_boxplot(aes(Species, value, colour=Species), show.legend=FALSE) +
  facet_grid(variable ~ .) +
  theme_bw()

If you want each plot in a separate panel, you can do this:

ggplot(melt(iris[,-1], id.var="Species")) +
  geom_boxplot(aes("", value, colour=Species), width=0.5, show.legend=FALSE) +
  facet_grid(variable ~ Species) +
  theme_bw() +
  theme(axis.ticks.x=element_blank()) +
  labs(x="")