Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Add subgroup labels to a jitter plot in ggplot2

Tags:

r

label

ggplot2

I have a nearly-boxplot like jitter-plot:

dt <- rbind(se,cb,cb.se)
qplot(ds, size, data=dt, geom="jitter", colour=root, facets = test ~ .)

plot

I'd love to put a summary label for each group in the middle of the plot - for example the size totals here:

 aggregate(list(size=dt$size), list(dt$ds, dt$test), sum)

   Group.1  Group.2   size
1     b217       se   9847
2      c10       se  97296
3     c613       se  21633
4       c7       se 207540
...

I've tried using + geom_text(aes(x=ds, y=128, label=sum(size)), size=2) to add labels, but I get the same label on each position - how can I get the sum of just that section of data?

Edit: Here's where I'm at now - maybe I'm just going in the wrong direction

data <- rbind(se,cb,cb.se)
labels <-ddply(data, c("ds", "test"), function(df) sum(df$size))
ggplot(data=data, aes(x=ds)) +
  geom_jitter(aes(y=size, colour=root)) +
  geom_text(data=labels, aes(x=ds, y=600, label=V1), size=3) +
  facet_wrap(test ~ .)

This code doesn't work - I get an undefined columns selected error... somewhere. Maybe it's because of the multiple data= sections?

like image 762
Thelema Avatar asked Jan 22 '10 19:01

Thelema


2 Answers

Since you don't provide sample data, I shall demonstrate a solution using random data.

set.seed(1)
n <- 100
dat <- data.frame(
    ds = sample(paste("x", 1:8, sep=""), n, replace=TRUE),
    size = runif(n, 0, 250),
    root = sample(c(TRUE, FALSE), n, replace=TRUE),
    test = sample(c("se", "cb", "cb.se"), n, replace=TRUE) 
)


head(dat)
  ds      size  root  test
1 x3 163.68098  TRUE cb.se
2 x3  88.29932  TRUE    se
3 x5  67.56504 FALSE    cb
4 x8 248.17102  TRUE    cb
5 x2 158.37332  TRUE    cb
6 x8  53.30203 FALSE cb.se

p <- ggplot(dat, aes(x=ds, y=size)) + 
  geom_jitter(aes(colour=root)) + 
  facet_grid(test~.) 

Create the data frame containing label data. Note the use of summarize. This tells ddply to create a new column to the data.frame

labels <- ddply(dat, .(ds, test), summarize, size=round(sum(size), 0))
head(labels)
  ds  test size
1 x1    cb  193
2 x1 cb.se  615
3 x1    se  274
4 x2    cb  272
5 x2 cb.se  341
6 x2    se 1012

p + geom_text(aes(x=ds, label=size, y=128), data=labels, size=2) 

enter image description here

like image 101
Andrie Avatar answered Nov 20 '22 16:11

Andrie


Take a look here. It may be helpful Adding direct labels to ggplot2 and lattice plots

like image 26
George Dontas Avatar answered Nov 20 '22 17:11

George Dontas