[R] standardize ggplot and lattice themes

[baptiste auguie]

Dear list,


I'm writing a long document (thesis) and as much as I would like to  
use only ggplot2 for the graphics, some features are still a bit  
undocumented so I often end up choosing either ggplot2, lattice, or  
base plots (which i know better) depending on the particular graph to  
produce. With the default settings, this does not make for a  
consistent look I would like to see throughout the document.
I really like most default settings in ggplot's theme (grid, labels  
spacing, background, ...), and I would like to produce a similar theme  
for use with lattice. I've made such an attempt in base graphics – see  
beow, it's just silly – but I think i could get away with using only  
lattice and ggplot2.

I don't know lattice well (just started using it, really), so I'm sort  
of lost in the jungle of parameters. Can anyone help me set up a  
lattice theme for the following example?

Best wishes,

baptiste

> library(ggplot2)
> library(lattice)
>
> # example data
> x <- seq(0, 10, len = 100)
> y1 <- jitter(sin(x), 1000)
> y2 <- 0.5*jitter(cos(x), 1000)
>
> # custom colors
> greyDark <- grey(0.5)
> greyLight <- grey(0.9)
> myColors <- c( "#E41A1C", "#377EB8")
> palette(myColors)
>
> # here is the ggplot2 version
> df <- melt(data.frame(x=x, one=y1, two=y2), id="x")
> p <- qplot(x,value, data=df,colour=variable, linetype=variable,  
> main="ggplot2 (almost) defaults")
> p <- p + scale_colour_manual(values = myColors)
> print(p)
>
> # lattice version
>  strip.background <- trellis.par.get("strip.background")
> background <- trellis.par.get("background")
>  plot.symbol <- trellis.par.get("plot.symbol")
>  trellis.par.set(strip.background = list(col = grey(7:1/8)))
>  trellis.par.set(plot.symbol = list(col = myColors, pch=16))
>  trellis.par.set(background = list(col = greyLight)) # this does not  
> clip to the plot region ...
>
> p2 <- xyplot(value ~ x, data=df)
> print(p2)
>
> # example using base graphics
> old.par <- par()
> par(cex=1, bty = "n",fg = greyDark, col.lab = "black",
> xpd = FALSE, mar = old.par$mar + c(-2,-1,-1,3), mgp=c(1.8, 0.5, 0),  
> col="black")
>
> plot(x, y1, new=TRUE, t="n") # plots nothing, needed to find the  
> dimensions
>
> lims <- par("usr")
> subGrid1 <- axTicks(1) + mean(diff(axTicks(1)))/2 # position of the  
> grid sub-divisions
> subGrid2 <- axTicks(2) + mean(diff(axTicks(2)))/2
>
> plot(x, y1, col=1, xlab = "x",  ylab = "value", xaxt = "n", yaxt =  
> "n", pch=16, cex=0.8,
> 	panel.first = {
> 	rect(lims[1], lims[3], lims[2], lims[4],bord = NA,  col =  
> greyLight); # grey background
> 	segments(axTicks(1),lims[3], axTicks(1), lims[4], col = "white" ,  
> lwd=1.2); # main grid
> 	segments(lims[2], axTicks(2),lims[3], axTicks(2), col = "white" ,  
> lwd=1.2);
> 	segments(subGrid1,lims[3], subGrid1, lims[4], col = "white" ,  
> lwd=0.5); # secondary grid
> 	segments(lims[2], subGrid2,lims[3], subGrid2, col = "white" ,  
> lwd=0.5);
> 	axis(1, lty = "solid", lwd = 1, col = greyDark, col.axis =  
> greyDark, tcl=-0.4, cex.axis = 0.8); # axis
> 	axis(2, lty = "solid", lwd = 1, col = greyDark, col.axis =  
> greyDark, las=1, tcl=-0.4, cex.axis = 0.8);})
>
> par(bty="o")
> box(col="white", lwd=3) # draws in white over the axes
>
> points(x, y2, col = 2, cex=0.8, pch=16) # some more plotting as usual
>
> par(xpd = TRUE) # legend is outside
> legend(1.1*max(x), mean(y1), pch=16, col=1:2, c("one", "two"),  
> bty="n", title="variable")
> title(main = "ggplot theme with base graphics")
>
> par(old.par)

_____________________________

Baptiste Auguié

School of Physics
University of Exeter
Stocker Road,
Exeter, Devon,
EX4 4QL, UK

Phone: +44 1392 264187

http://newton.ex.ac.uk/research/emag