[R] Factor level coloring in trellis plot
Deepayan Sarkar
deepayan at stat.wisc.edu
Thu Feb 17 15:52:28 CET 2005
On Thursday 17 February 2005 04:39, T.A.Wassenaar wrote:
> Hi :)
>
> Was just wondering whether someone could help me with
> adjustments to trellis plots (parallel).
>
> I've got two way multivariate data. I want to make
> parallel plots for one of the factors, and want to color
> the lines according to the other factor. The first thing I
> manage, but with the other I'm lost :( Can only change the
> overall color.
>
> This is basically how far I get:
>
> parallel(~data | factor, layout=c(4,1))
>
> Any hints will be greatly appreciated.
panel.parallel doesn't handle groups, so you'll need to write a panel
function that does. This is easy enough, the current panel function
needs only a few additions:
panel.parallel.new <-
function(z, subscripts,
groups = NULL,
col=superpose.line$col,
lwd=superpose.line$lwd,
lty=superpose.line$lty, ...)
{
superpose.line <- trellis.par.get("superpose.line")
reference.line <- trellis.par.get("reference.line")
n.r <- ncol(z)
n.c <- length(subscripts)
if (is.null(groups)) {
col <- rep(col, length=n.c)
lty <- rep(lty, length=n.c)
lwd <- rep(lwd, length=n.c)
}
else
{
gnum <- as.integer(as.factor(groups))
n.g <- length(unique(gnum))
col <- rep(col, length=n.g)
lty <- rep(lty, length=n.g)
lwd <- rep(lwd, length=n.g)
}
llim <- numeric(n.r)
ulim <- numeric(n.r)
dif <- numeric(n.r)
if (n.r > 0)
for(i in 1:n.r) {
grid.lines(x = c(0,1), y = c(i,i),
default.units = "native",
gp = gpar(col = reference.line$col,
lwd = reference.line$lwd,
lty = reference.line$lty))
llim[i] <- range(as.numeric(z[,i]))[1]
ulim[i] <- range(as.numeric(z[,i]))[2]
dif[i] <- ulim[i] - llim[i]
}
if (is.null(groups))
for (i in seq(along=subscripts))
{
x <- (as.numeric(z[subscripts[i],,])-llim)/dif
grid.lines(x = x,
y = 1:n.r,
gp = gpar(col=col[i], lty=lty[i], lwd=lwd[i]),
default.units="native")
}
else
for (i in seq(along=subscripts))
{
x <- (as.numeric(z[subscripts[i],,])-llim)/dif
grid.lines(x = x,
y = 1:n.r,
gp =
gpar(col=col[gnum[subscripts[i]]],
lty=lty[gnum[subscripts[i]]],
lwd=lwd[gnum[subscripts[i]]]),
default.units="native")
}
invisible()
}
With this, you should get what you want with
library(grid)
library(lattice)
parallel(~iris[1:4], iris,
panel = panel.parallel.new,
groups = Species)
I'll include this in the next release.
Deepayan
More information about the R-help
mailing list