4

I was wondering if it was possible to produce a set of boxplots similar to that produced by this nested loop combinations using an apply function.

It may not be possible/necessary but I thought it should be possible, I just cant wrap my head around how to do it.

I need to be able to plot this to see how 100s of factors compare in respect to one variable (mtcars$mpg)

head(mtcars)

for (i in 8:11) {
    for (j in 8:11) {

        if (i != j) {

            title = paste(names(mtcars)[i], names(mtcars)[j], 
                sep = "/")

            p <- ggplot(mtcars, aes(interaction(mtcars[,i], mtcars[, j]), mpg, fill = factor(mtcars[,i]))) + geom_boxplot(alpha = I(0.7))
            p <- p + ggtitle(title) + scale_fill_hue()


        } else {

            title = paste(names(mtcars)[i])

            p <- ggplot(mtcars, aes(factor(mtcars[,i]), mpg, fill = factor(mtcars[, i]))) + geom_boxplot(alpha = I(0.7))
            p <- p + ggtitle(title) + scale_fill_hue()


        }

        print(p)
    }
}
aeongrail
  • 1,304
  • 1
  • 13
  • 26

1 Answers1

3

Put the if block in a function:

plotGG <- function(i,j) 
  {
  if (i != j) { ... } else{ ... }
 }

Then call it:

mapply(plotGG,8:11,8:11)

And it works.

Your code will not work due to a scoping issue with ggplot. But you can view the solution here: Local Variables Within aes

EDIT: You can finish wrap it as you want:

multiPlotGG <- function(l1,l2) {
 mapply(plotGG,rep(l1,each = length(l2)),rep(l2,length(l1)))
}
multiPlotGG(8:11,8:11)
Community
  • 1
  • 1
Shambho
  • 3,250
  • 1
  • 24
  • 37