The hottest commodity in the sports analytics world right now is player tracking data. It’s called different things in different sports (e.g. “Next Gen Stats” in the NFL), but it all boils down to somehow measuring exactly where every player is on the field/pitch/court in very short intervals (e.g. 10 times per second).
These data sets are extremely rich, complex, and big. For example, in a 7-second American football play, you would have 22 players x 10 observations per second x 7 seconds = 1,540 observations of several metrics (at a minimum, X and Y coordinates on the field). Just wrangling this data into analyzable shape is an enormous challenge, but I’m not writing about that today because I’m not an expert in data engineering.
Instead, let’s say you have a bunch of this data cleaned, imported, and ready to analyze. What to do next can still feel overwhelming. Where do you even start? My goal with this post is to provide a unified, cross-sport list of high-level options for things you could calculate with tracking data based on the work that I’ve seen; I want to make the problem of what to do next less abstract.
A couple caveats: first, I come at this from a sports science/player performance/injury perspective, rather than from fan engagement or in-game strategy. This is supposed to be a living document (LAST UPDATE: February 12, 2020), and I’m hoping other people will help me flesh out this list with things I’ve missed to make it more comprehensive. But cut me a little slack if I miss something obvious from outside my expertise – we all need diverse teams to do great work. Second, this post is designed to be a list of metrics you could calculate from the data, not questions you can answer with it. Hopefully this list of metrics a.) helps break the logjam of “Oh God where do I start with all this?” and b.) helps suggest interesting questions you could investigate with some of them. Third, there are a virtually infinite number of ways you could combine and tweak all these metrics – if there’s a popular one of these you like that I didn’t specifically include I’m happy to add it to the list, but I wanted to start with the overarching concepts. A detailed, near-infinite list might be useful for some people, but it wasn’t my goal here. Consider this more an attempt at a taxonomy.
*Deep breath* OK, let’s go.