The newly founded Sports Analysis Collective at UW-Madison wanted to create a predictive model for NBA games so that we could publish articles about our findings. We wanted to see which statistics are the most accurate indicators of winning in the NBA.
What it does
The predictive algorithm allows users to choose one home and one away team as a game matchup. It takes into account various team statistics, including offensive efficiency, the results of the previous five games played by each team, and more. Next, the algorithm calculates a projected score of the game.
How I built it
We built a web scraper for the back end and ran the analytics using Java, our gui was also built using java.
Challenges I ran into
Difficult to manage the amount of time each scraping test took.
Accomplishments that I'm proud of
Our web scraper worked very successfully, unfortunately our model and UI were not up to par.
What I learned
Learned how to scrape large amounts of data from the web using jsoup.
What's next for NBA Predictive Model
We hope to improve the model's user interface to and expand on the statistics taken into account for each matchup. Later it will be published on the organization's website and used in articles.