Spotting Differences - Data Points: Visualization That Means Something - page 240

FIGurE 7-10 Default Chernoff Faces

So for example, the height of a face represents the number of games
played, and the height of a mouth represents field goals made per game.
This is kind of useless as it is because you don't have any names to the
faces, but you can see that the first few have more well-rounded gameplay
than the others, whereas player 7, for instance, has relatively wide hair,
which corresponds to three-pointers made.

When you have a
lot of individuals, it
can be useful to
cluster by cat-
egory so that the
faces are easier
to scan. For this
example, you
could separate
faces by position:
guards, forwards,
and centers.

Use the labels argument in faces() to add names to the faces, as shown in
F igur e 7-11.
faces(bball[,2:16], labels=bball$Name)

That's better. Now you can see which face corresponds to which player.
To identify the point guards, you can start with say, Chris Paul, and look
for similar faces such as the one for Devin Harris or Deron Williams.
Chauncey Billups, in the bottom-right corner is also a point guard, but the