r- axis label in image

Question

I need to graph a metric by spatial position in a call center. I wrote in R a tiny example:

tt<-data.frame(a1=c(0.4,.5,.5,.7),a2=c(.5,.6,.7,.8), a3=c(.8,.7,.9,.8))
row.names(tt)<-paste("L", 1:4, sep='')
tt<-as.matrix(tt)
tt

So my matrix is:

> tt
    a1  a2  a3
L1 0.4 0.5 0.8
L2 0.5 0.6 0.7
L3 0.5 0.7 0.9
L4 0.7 0.8 0.8

I tried:

palette <- colorRampPalette(c('#f0f3ff','#0033BB'))(256)
library(fields)
image.plot(t(tt[rev(order(row.names(tt))),]),col = palette, axes=F ,
       lab.breaks=NULL)

I had to transpose and reorder the matrix because I wanted the way you read it in the table.

So I got:

I need to add next to each square the row and column names. For example, top left squared should have "L1" at the left and "a1" at the top.

Also I would like to add the values in each square.

I tried axis() but I got the wrong result. I'm pretty new in doing graphs in R, so any help would be appreciated.

Answer 1

The axis command doesnt seem to plot the labels outside the image, even when using the outer argument.

It does work however, if you use the image function (with add=TRUE) after image.plot

image.plot(t(tt[rev(order(row.names(tt))),]),col = palette, axes=FALSE, 
                                                          lab.breaks=NULL)

# axis labels
image(t(tt[rev(order(row.names(tt))),]), col = palette, axes=FALSE, add=TRUE)
axis(3, at=seq(0,1, length=3), labels=colnames(tt), lwd=0, pos=1.15)
axis(2, at=seq(1,0, length=4), labels=rownames(tt), lwd=0, pos=-0.25)

# add values
e <- expand.grid(seq(0,1, length=3), seq(1,0, length=4))
text(e, labels=t(tt))

Answer 2

I think you'll be much happier using ggplot -- it makes this sort of thing easy, much less error-prone, and plot code is more readable. For this, you'll want to keep your data in a dataframe, and melt it to be in "long form" (here I use melt from the reshape2 package, but you could also just set your dataframe up in this form to start with). Try this:

library(ggplot2)
library(reshape2)

tt<-data.frame(a1=c(0.4,.5,.5,.7),a2=c(.5,.6,.7,.8), a3=c(.8,.7,.9,.8))
tt$row <- paste("L", 1:4, sep='')
tt_melt <- melt(tt)

ggplot(data=tt_melt,
       aes(x=variable, y=row, fill=value)) + geom_tile() + 
       geom_text(aes(label=value), color='white') + theme_bw()

ggplot also allows you to control the color scale if you wish. If you're going to be plotting in R, it's well worth investing a couple hours learning ggplot!