On June 19, the NBA finals concluded. The Cleveland Cavaliers and Golden State Warriors played a historically great series, culminating in a must-see seventh game that more than 30 million Americans watched. And what a game it was. By some advanced measures, we've never seen two better teams meet on the world's largest basketball stage.
I've been thinking quite a bit about basketball in the context of this month's theme: data preparation for analytics. Specifically, the last few years have given rise to advanced analytics in many sports, especially basketball.
Sure, for years stat-heads have crunched their own numbers. (Fantasy sports and gambling have served as accelerants here.) Even the biggest geeks, however, could not accurately capture a great deal of valuable game- and player-specific data. For instance, Clippers' center / slam-dunk machine DeAndre Jordan routinely shoots a high field-goal percentage (except from the free throw line).
As any casual fan knows, not all shot attempts are good ones. There are "bad two's" and "good three's." What if in-game data could shed light on the the quality of each player's shots?
utilizes a six-camera system installed in basketball arenas to track the real-time positions of players and the ball 25 times per second. Utilizing this tracking data, STATS is able to create a wealth of innovative statistics based on speed, distance, player separation and ball possession.
In a word, wow. Let's get specific though.
Over the course of his career, Kobe Bryant made his fair share of buzzer beaters, but any professional can hit a spectacular last-second shot or two if given enough opportunities. More compelling questions include the following:
- How often do players take what we consider to be good shots?
- Which players make a higher percentage of clutch shots?
- Which areas of the court are most conducive to good shots?
- Which players create high-quality shots for their teammates?
Try answering these questions objectively (re: sans data) a decade ago.
Looking at new data sources, we can answer that question. (Trust me, I know a thing or six about bad shots. See my embarrassing stats from last year's three-on-three match with my college friends. A few years ago, my friends created the category of uglies with me in mind, but I digress.)
Now, no one ever needed anything as advanced as SportVU® to critique my basketball skills. The same cannot be said, however, for NBA players – especially in the context of impending free agency, a complicated salary cap and luxury taxes.
Simon Says: New technologies make analytics better.
Sensors and wearable technology aren't always accurate. (Exhibit A: the Fitbit.) Limitations aside, the NBA demonstrates that recent innovations have in many cases reduced or even eliminated the need for manual data preparation. In the hoops world, SportVU® and their ilk have increased data speed and accuracy, never mind allowing for previously impossible analytics.
If you're not the general manager of the Lakers, don't fret. With the Internet of Things, expect this trend to explode in the coming years.
What say you?