This repo contains a machine-learning model that predicts the box scores of future NBA games based on data from seasons 2016-2024
- Web Scraping
- Uses BeautifulSoup and Playwright to scrap data from past NBA games
- Gets the data from basketball-reference.com
- Parses the data into desired format
- Extract game statistics and create new features needed for machine learning later
- Output nba_games.csv: a CSV file with all data parsed and ready
- Machine Learning
- Uses RidgeClassifier to make predictions
- Trains the model with the data from the first few seasons in nba_games.csv
- Checks accuracy by predicting box scores of games in the latter half of nba_games.csv