Lamelo's Balls

Landing screen
Search results for Jalen Brunson

Inspiration

Bassgedbol :)

What it does

Predicts an NBA player's statline for their upcoming game, when given the opponent team's name and the player's name.

How we built it

We wrote Python scripts to scrape online basketball statistics websites to retrieve both player and team stats. We then treated this data for ease of future handling, using both Tcl and Python.

The project was built with a random forest regressor, trained on the individual stats of the players and their opposing teams' stats for the previous 10 games of all players in the NBA.

We also made a "fantasy points" field calculated from the other

Challenges we ran into

It don't work too good.

Accomplishments that we're proud of

We achieved a 75% accuracy (where predictions are considered correct if they're within 3 of the correct result), where ESPN has ~80% accuracy in their projections, within a greater margin of error.