MSDS621 Final Project Group Members: Chris Olley, Connor Swanson, Anish Dalal, Jon-Ross Presta, & Sankeerti Haniyur
For our final project, we wanted to create a machine learning algorithm that could predict Kobe Bryant's shot success. Using this dataset, we were able to create Random Forest and Gradient Boosting models with log loss of 0.6103 and 0.6064 respectively.
Please use the labeled folders in this repository to find our code for our expoloratory data analysis, fitting multiple models, feature engineering, model evaluation, and visualization of our data.
- Data source: https://www.kaggle.com/c/kobe-bryant-shot-selection/data
- Final presentation: https://github.com/anishpdalal/mamba-mentality/blob/master/FinalPresentation.ipynb
- Final Modeling Approach: https://github.com/anishpdalal/mamba-mentality/blob/master/FINAL_models.ipynb
- Contains model comparison via cross-validation results
- Tests for ensuring cross-validation behaved as expected
- EDA and Data Visualization: https://github.com/anishpdalal/mamba-mentality/blob/master/Final_EDA.ipynb
- Contains code for visualizations included in the final presentation
- To see other plots, please visit the eda__data_viz folder
- First project check-in: https://github.com/anishpdalal/mamba-mentality/blob/master/final_project_checkin_template.ipynb