What Geography Can Teach Us About Basketball

The problem with the date (and you can view the maps at the original link) is that it will not work in the real world because basketball is a fluid game. Players already know where there best chance of hitting shots is - the question is, will the defender allow them to get to the spot on the floor where they can make the shot, and once they get to that spot, will the defender put a hand in their face or let them shoot uncontested?

Still, it's amusing.

From Slate: What Geography Can Teach Us About Basketball

The annual Sloan Sports Analytics Conference, created in 2006, has become something like Bonnaroo for sports nerds. And if there was a breakout star at this year's gathering, held at MIT this past weekend, it may have been Kirk Goldsberry, an assistant professor of geography at Michigan State (and currently a visiting scholar at Harvard). At Sloan, Goldsberry—whose dissertation "investigated real-time traffic maps" and who has also used geography to examine "access to nutritious foods in urban areas"—considered the ways that sophisticated statistical mapping can illuminate the game of basketball, in a paper called "Court Vision: New Visual and Spatial Analytics for the NBA."

You only have to glance at the maps Goldsberry produced to know that stat-friendly teams will pounce on these things. As the New York Times basketball blog Off the Dribble noted over the weekend, "about a third of the league’s arenas have recently installed camera systems that capture and log the position of every player on the court 25 times a second." As a result, many teams now have incredible amounts of data they can visualize in some of the ways Goldsberry suggests.

For the map above [not included here], for instance, Goldsberry divided roughly half an NBA court (from the baseline to just past the 3-point line) into 1,284 "shooting cells." Then he plotted every shot taken in an NBA game from 2006 to 2011, and color-coded the results. The areas which yielded the most points per shot appear near the red end of the color spectrum; those that yielded the fewest are at the blue end.

If you've read anything about scoring efficiency in basketball, the resulting image will not surprise you (though its elegance is striking). But it conveys far more quickly and powerfully than a set of numbers can what kind of shot distribution an NBA team should be going for, generally speaking.

And this kind of visualization becomes even more useful when you switch from the general to the specific. Goldsberry sent me a map he created tracking the shots of my favorite player, Boston Celtics point guard Rajon Rondo. The hexagons that represent each shooting area are larger where Rondo shoots a lot, smaller where he doesn't. And while the image echoes the league-wide results, it also reveals some personal idiosyncrasies—such as the fact that Rondo, a poor outside shooter for the most part, actually has one sweet spot when it comes to three-pointers. [Map of Rondo not included.]

From a map like this, a player can quickly learn both where he should try to shoot during a game (the red spots) and where he should shoot during practice (the blue spots). A coach, meanwhile, could layer the equivalent map for each one of his players on top of one another and find in the visual data inspiration for new plays that lead each man to one of his sweet spots. And for the mere fan, such maps can not only lead to a greater understanding of the game, but also provide at least a hint of the aesthetic pleasure that makes basketball enjoyable in the first place. As Michael Scott might say, win-win-win.