I've created a heatmap in R using the heatmap2 from the gplots package and I am having awful trouble trying to format the image for use in a report.
The image represents the eigenvalues of a small correlation matrix of wavelet coefficients and I wish to represent the largest eigenvalues with a strong colour such as red and smallest using a soft yellow or similar. The colour palette isn't so much the issue it's the representation of the colours. Currently yellow represents the largest eigenvalue so I wish to reverse the order if possible.
Also is it possible to rotate the legend similar to the first image below, I've checked the help for such info but all I can find is row and column label rotation?
Since my data represents a week's worth of data taken every hour, how can I change the x-axis values to show the following 12/24hr vector: 12,24,36,48,60,72,84,96,108,120,132,144,156,168. I tried setting this using cexCol but I got an error about cex.axis having the wrong length and I don't see this as an argument in heatmap2.
par(mfrow=c(1,1))
heatmap.2(eigenvalsCombined,
trace = "none",
dendrogram = "none",
Rowv = NULL,
Colv = NULL,
density.info = "none",
margin = c(5,7),
main = expression(paste("Heatmap of Largest Eigenvalues ",
lambda[1],
" Across 7 Wavelet Scales")),
xlab = "Time Index (hours)",
key = TRUE,
lmat = rbind(c(2,3),c(4,1)),
key.title = NA,
key.xlab = "Eigenvalue Magnitude")
As you can see from my image below the main title is also chopped off, I have tried playing with the outer and inner margins but this doesn't have any impact either. When I use the zoom feature in the plotter and blow up the image the text appears, does this mean my formatting is correct?
For those interested in the data, I have posted a copy of the dput(eigenvalsCombined) to google drive for download.