This was part of the UChicago PATHWAYS IN DATA SCIENCE Summer School project in July-August 2021. The summer school was taught by PHILLIP LO and ABBY STEVENS, as an intro to college level machine learning. The final project was to come up with a model using one of the popular ML algorithm: linear regression, KNN, decision tree... on an online dataset. I employed hollinger's team and game datasets to train a random forest on three season of NBA games: 18-19, 19-20 and 20-21.

Files:

• game_results_18.csv : data file for each game in the 18-19 season, with home_team_ID, away_team_ID and 1/0 for home win/loss
• game_results_19.csv : data file for each game in the 19-20 season, with home_team_ID, away_team_ID and 1/0 for home win/loss
• game_results_20.csv : data file for each game in the 20-21 season, with home_team_ID, away_team_ID and 1/0 for home win/loss
• team18.csv: data containing 6 features measuring the performance of every team in that season, and we will use those as predictive variable.
• NBA_randomforest.ipynb : Jupyter notebook with python implementation of this project.

Features:

• PACE : Pace Factor - the number of possessions a team uses per game.
• AST : Assist Ratio - the percentage of a team's possessions that ends in an assist.ss
• TO : Turnover Ratio - the percentage of a team's possessions that end in a turnover.
• REBR : Rebound Rate - the percentage of missed shots that a team rebounds.
• OFF EFF: Offensive Efficiency - the number of points a team scores per 100 possessions.
• DEF EFF: Defensive Efficiency - the number of points a team allows per 100 possessions.