R: how to change lattice (levelplot) color theme?

Question

The default theme on my installation is something which maps the values to pink and cyan. How to change it for example to a gray scale theme?

Answer 1

You may also want to consider using panel.levelplot.raster, as it reduces the figure size sizably, as I recently found out. Combining this tip with changing trellis settings in general, here are a following examples:

trellis.par.set(regions=list(col=topo.colors(100)))
levelplot(volcano, panel = panel.levelplot.raster)

levelplot(volcano, panel = panel.levelplot.raster,
          par.settings=list(regions=list(col=topo.colors(100))))

The second method is deceptive because the trellis settings are in fact being changed globally. I didn't know about the col.regions argument though - that's pretty nice as it seems to change the color theme locally.

Answer 2

There are many ways to accomplish your request. The simplest is:

trellis.device(color = FALSE)

another is

ltheme <- canonical.theme(color = FALSE) ## in-built B&W theme 
ltheme$strip.background$col <- "transparent" ## change strip bg 
lattice.options(default.theme = ltheme) ## set as default 

See this mail archive item: Re: [R] specify lattice black-and-white theme for more info.

Answer 3

You can use

library(lattice)
lattice.options(default.theme = standard.theme(color = FALSE))

which turns on black-and-white suitable for printing. I've also played with things like

sb <- trellis.par.get("strip.background") 
sb[["col"]][1] <- "lightgray"
trellis.par.set("strip.background", sb) 

which changes just the header background.

And I thought people only asked ggplot2 questions here :) Nice to see some lattice for a change.

Answer 4

The only one that worked for me is:

>trellis.par.set(canonical.theme(color = FALSE)) 

Answer 5

Thanks for the answers guys! It also helped me to find more information on the subject. I learned that I can control the scales of gray using for example the following:

levelplot(my_var, col.regions = gray(0:100/100))

which gives me 100 shades of gray from black (0) to white (1).

I'm using the function to plot gray scale images (photos) which I've pre-processed to a double matrix. I don't know if it's the best possible approach, but so far it works and I believe it gives me more options for graphing than the basic displaying options in the EBImage and rimage libraries. I don't know how I'd alter the palette to match displaying color images, but I'm glad I didn't have to do that so far...