Different Plottypes in facet_grid - r

Question

once again Im confronted with a complicated ggplot. I want to plot different plottypes within one plot using facet grid.
I hope I can make my point clear using the following example:
I want to produce a plot similar to the first picture but the upper plot should look like the second picture.
I already found the trick using the subset function but I can't add vertical lines to only one plot let alone two or three (or specify the color).
CODE:
a <- rnorm(100)
b <- rnorm(100,8,1)
c <- rep(c(0,1),50)
dfr <- data.frame(a=a,b=b,c=c,d=seq(1:100))
dfr_melt <- melt(dfr,id.vars="d")
#I want only two grids, not three
ggplot(dfr_melt,aes(x=d,y=value)) + facet_grid(variable~.,scales="free")+
geom_line(subset=.(variable=="a")) + geom_line(subset=.(variable=="b"))
#Upper plot should look like this
ggplot(dfr,aes(x=d,y=a)) + geom_line() + geom_line(aes(y=c,color="c"))+
geom_hline(aes(yintercept=1),linetype="dashed")+
geom_hline(aes(yintercept=-2),linetype="dashed")

Answer 1

If I understand your question correctly, you just need to a variable column to dfr in order to allow the faceting to work:
dfr$variable = "a"
ggplot(subset(dfr_melt, variable=="a"),aes(x=d,y=value)) +
facet_grid(variable~.,scales="free")+
geom_line(data=subset(dfr_melt,variable=="a")) +
geom_line(data=subset(dfr_melt, variable=="b")) +
geom_line(data=dfr, aes(y=c, colour=factor(c))) +
geom_hline(aes(yintercept=1),linetype="dashed")+
geom_hline(aes(yintercept=-2),linetype="dashed")
Notice that my plot doesn't have the zig-zig line, this is because I changed:
#This is almost certainly not what you want
geom_line(data=dfr, aes(y=c, colour="c"))
to
#I made c a factor since it only takes the values 0 or 1
geom_line(data=dfr, aes(y=c, colour=factor(c)))
##Alternatively, you could have
geom_line(data=dfr, aes(y=c), colour="red") #or
geom_line(data=dfr, aes(y=c, colour=c)) #or

Answer 2

To my knowledge, you can't put multiple plot types in a single plot using facet.grid(). Your two options, as far as I can see, are
to put empty data in the first facet, so the lines are 'there' but not displayed, or
to combine multiple plots into one using viewports.
I think the second solution is more general, so that's what I did:
#name each of your plots
p2 <- ggplot(subset(dfr_melt, variable=="a"),aes(x=d,y=value)) + facet_grid(variable~.,scales="free")+
geom_line(subset=.(variable=="a")) + geom_line(subset=.(variable=="b"))
#Upper plot should look like this
p1 <- ggplot(dfr,aes(x=d,y=a)) + geom_line() + geom_line(aes(y=c,color="c"))+
geom_hline(aes(yintercept=1),linetype="dashed")+
geom_hline(aes(yintercept=-2),linetype="dashed")
#From Wickham ggplot2, p154
vplayout <- function(x,y) {
viewport(layout.pos.row=x, layout.pos.col=y)
}
require(grid)
png("myplot.png", width = 600, height = 300) #or use a different device, e.g. quartz for onscreen display on a mac
grid.newpage()
pushViewport(viewport(layout=grid.layout(2, 1)))
print(p1, vp=vplayout(1, 1))
print(p2, vp=vplayout(2, 1))
dev.off()
You might need to fiddle a bit to get them to line up exactly right. Turning off the faceting on the upper plot, and moving the legend on the lower plot to the bottom, should do the trick.