About

I created this site to help fill in some gaps in the basketball analytics landscape. I've attempted to solve several hurdles that currently exist:

• The Play-by-play data provided by the NBA is useful but doesn't come ready made. It contains event-level data, but some important information information needs to be parsed out by the analyst. The most important example is that stat events aren't grouped by possession out of the box. The combined_pbp dataset solves this by giving stat events a possession id.
• As far as I know, there doesn't exist a publicly available win probability model. inpredictable.com is currently the go to for win probability, but that data isn't available in table form. The possessions dataset from this site's API includes the home team's win probability at the start of each possession.
• One of the most important metrics for basketball research is Regularized Adjusted Plus-Minus (RAPM). And while different versions can be found, there isn't one place for historic and current season RAPM that is easily accessible. I have made both 1 and 3 year RAPM available on the site and its API (5-year RAPM is coming soon).

While I'm not the first to tackle any of these problems, my goal is to solve them in the most accessible way possible.