Just an avid (Miami Heat) fan playing around with NBA data in my spare time
- [DONE] Different combos of (PTS, AST, REB, STL, BLK) vs (Point Differential and wins)
- [DONE] Clusters of players based on metric (FG%, STL+BLK, etc..) (Hierarchial to get some sense of tiers of players, Categorical for player types: shooter, defender, 2-way, 3&D, athletic, etc)
- Determine and predict who will win a game a) [DONE] Compare teams in a game based on different recorded stat accumulation in quarter-bucketized metrics b) Predict outcome of the game based on difference of the stats in the comparision
- [Can't do this no way to get info] Playoff vs Regular season comparisons of: a) Homecourt Advantage b) player stats -> Use the difference to get tiers of players who "step it up", stay the same, underperform (choke, etc) c) point differential -> Does it actually get harder to win in the playoffs?
- [DONE] Determine IMPORTANT points ("Takeover" moments or "heat-check" shots) based on game win probability find inflection points to point to a play or event in game that is the takeover moment
- Important shots in a game: Maybe shot chart vs distance from the rim (usually players pull from deep when they are feeling it)
- [IN PROGRESS] Figure out quality of team (Offensive efficiency) based on team shot chart efficiency
- Player shot chart weighted into how deep into a game, and then extend it to how deep into a season
- Use with AWS services Cloudfront, APIGateway, Lambda, DynamoDB, Cloudwatch to continuosly update
- Integrate with social media (e.g twitter API)
- Add a mobile app/website aspect to this (Angular/ReactJS)
Utilizing API's written against nba.com found here: https://github.com/swar/nba_api
Peter Beshai ----> https://peterbeshai.com/