March Madness Machine Learning Competition 2026

Technologies

1. NumPy for easy array manipulation
2. pandas for visualization purposes
3. scikit-learn for model evaluation and prediction
4. sqlite for manipulating the .csv data files

Results

1. Round of 64: 26 / 32, meaning 26 points
2. Round of 32: 11/16, meaning 22 points
3. Sweet 16: 3/8, meaning 12 points :(
4. Elite 8: 4/4, meaning 32 points
5. Final 4: 2/2, meaning 32 points
6. National Championship: 0/1, meaning 0 points

Notes

• I only used 11 seasons (from 2014 to 2025) to train and test my model, though there is data that goes back to the early 2000s. I believe CBB has changed a lot since then, so I wanted to focus on a more modern side of CBB.
• Information about the ML competition is here
• Information about the actual tournament is here
• My personal pick to win MM is Vanderbilt, mainly because I played against their star player Tyler Tanner in high school (damn, they lost 🙃).