Planning Hockey Careers With Python
summary
A deployed career-planning feature uses historical player paths from ~25,000 profiles, filters for same start age and league, builds nodes as age+league pairs and edges weighted by how many players took that step, converts weights to reciprocal distances and finds shortest-path on a directed weighted graph to suggest multiple routes and ppg targets; messy data and coverage issues leave only ~20% direct matches so the graph approach boosts coverage, implemented with DuckDB and NetworkX and served via FastAPI.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:31Speaker and company
- 6:10Career planning problems
- 9:48Theory and data modeling
- 13:47Predicting career paths
- 17:00Python implementation
- 23:52Summary
- 25:17Q&A
related talks
04mInteractive Sports Analytics
23mSport analysis with Python
32mCutting Edge Football Analytics in Python
35mBuilding a Player-Focused Data Team
29mFootball (soccer) data analysis: A pedagogic introduction
01h 33mBuilding an Interactive Network Graph to Understand Communities
41mEnhancing Event Analysis at Scale: Leveraging Tracking Data in Sports
41mThe evolution of basketball with data science