R - Correlation heatmap created with ggplot2: How can I flip the labels on the y-axis?

Question

I found this easy way to create a correlation heatmap using ggplot2:

data(attitude) 
library(ggplot2) 
library(reshape2) 
ggplot(melt(cor(attitude)), aes(Var1, Var2)) +   
     geom_tile(aes(fill = value))

However, this puts the variances on the antidiagonal of the matrix. I expected them on the diagonal. I tried to accomplish this using the rev() command:

ggplot(melt(cor(attitude)), aes(Var1, rev(Var2))) +
  geom_tile(aes(fill = value))

Which works fine as far as the colouring scheme inside the tiles is concerned. But the labels on the y-axis remain unchanged! What can I do about that? I'd rather not type in the correct order by hand, because my code should work for arbitrary data sets.

Answer 1

I've had this problem before, found an answer here: https://stackoverflow.com/a/14630937/4090947

Basicly:

data(attitude) 
library(ggplot2) 
library(reshape2) 

ggplot(melt(cor(attitude)), aes(Var1, ordered(Var2, levels =     rev(sort(unique(Var2)))))) +   
geom_tile(aes(fill = value))

This reverses your Y-axis

As your column names are discretes the key is to convert the strings (column names) to factors. Once you have factors, you can change the order.