This project builds a simple predictive model to estimate the probability of UVA Women’s Basketball reaching the Final Four.
Using key performance metrics:
- Offensive rebound rate
- Defensive rebound rate
- Turnover rate
- Free throw percentage
- Field goal percentage
I created a weighted scoring system to estimate matchup win probabilities.
- Calculated differences between UVA and opponent stats
- Applied weights to each feature
- Converted scores to probabilities using a logistic function
- Modeled sequential probabilities to estimate Final Four likelihood
- Sweet 16 vs TCU: ~19.7%
- Elite 8 vs South Carolina: ~13.9%
- Elite 8 vs Oklahoma: ~14.1%
- Final Four probability: ~2.8%
While UVA has competitive metrics, the compounded difficulty of sequential matchups significantly reduces Final Four probability. I still will be cheering for them all the way, GO HOOS!
- Python (pandas, numpy)
- Tableau (visualization)
Built as part of my MSDS journey at UVA
