Data driven plot names in data.table

Question

This is a personal project to learn the syntax of the data.table package. I am trying to use the data values to create multiple graphs and label each based on the by group value. For example, given the following data:

# Generate dummy data
require(data.table)

set.seed(222)
DT = data.table(grp=rep(c("a","b","c"),each=10), 
 x = rnorm(30, mean=5, sd=1), 
 y = rnorm(30, mean=8, sd=1))

setkey(DT, grp)

The data consists of random x and y values for 3 groups (a, b, and c). I can create a formatted plot of all values with the following code:

# Example of plotting all groups in one plot
require(ggplot2)

p <- ggplot(data=DT, aes(x = x, y = y)) + 
  aes(shape = factor(grp))+
  geom_point(aes(colour = factor(grp), shape = factor(grp)), size = 3) +
  labs(title = "Group: ALL")
p

This creates the following plot:

Instead I would like to create a separate plot for each by group, and change the plot title from “Group: ALL” to “Group: a”, “Group: b”, “Group: c”, etc. The documentation for data.table says:

.BY is a list containing a length 1 vector for each item in by. This can be useful when by is not known in advance. The by variables are also available to j directly by name; useful for example for titles of graphs if j is a plot command, or to branch with if()

That being said, I do not understand how to use .BY or .SD to create separate plots for each group. Your help is appreciated.

Answer 1

Here is the data.table solution, though again, not what I would recommend:

make_plot <- function(dat, grp.name) {
  print(
    ggplot(dat, aes(x=x, y=y)) + 
    geom_point() + labs(title=paste0("Group: ", grp.name$grp))
  )
  NULL
}    
DT[, make_plot(.SD, .BY), by=grp]

What you really should do for this particular application is what @dmartin recommends. At least, that's what I would do.