The idea behind Chernoff Faces it that you can take data and map it to facial features in order to visually present differences in the numbers. The theory is that humans easily recognize facial features and can notice changes fairly quickly. There’s a lot of debate as to how useful Chernoff Faces are, but they’re a fun diversion and, I think, an interesting take on displaying data sets.
What I’ve done for today’s ‘Wednesday Graph’ is to take 10 statistics — wOBA, K%, BB/K, AVG, OBP, ISO, Speed Score, BABIP, BB%, and FIP — and create Chernoff faces for every major league team. I’ll also note here, to avoid any confusion, that the BB% and K% statistics are based on offense and not pitching.
I would also like to send a huge thank you out to Flowing Data for putting up a tutorial on my request. FD has some of the best tutorials on the web for learning and using R and it’s a daily staple for graph junkies like myself. R is still very new to me and this post wouldn’t be possible without FD.
(Click to enlarge)
Be sure to consult the key above for what features are being displayed by which statistic. For example, the Diamondbacks’ hair is almost horn-like, representing their league worst FIP (4.82). Compared to the league’s best pitching (Padres, Braves) the D-Backs hair looks pretty goofy. We’ve mapped isolated-power (ISO) to the smiling trait. The more the face is smiling, the more power that team has hit for. The Mariners, Astros, and Royals have been a few of the most punch-less teams in baseball this year. Their individual ISOs of .106, .113, and .119 are extremely poor and they look almost depressed. On the other side of things, the Blue Jays (.203!), Red Sox (.185), and the Yankees (.174) boast some of the best power in baseball with their huge grins.
I’ll leave you to read over the graph some more, but if anyone has any questions, or suggestions, please post them in the comments section.