Revolutionizing Football: Using Machine Learning to Predict Future Performances for Quarterbacks

It is important to have a reliable application that can aid users of betting and fantasy football in which players they should put bet on, or choose for their fantasy teams. My motivation behind this project was to create something that could further betting and fantasy football, and even increase traction.

Football is a largely popular American sport, and within fans of the sport, betting and fantasy football are also very popular. Football is also very unpredictable. On a certain day, a player may perform better because of the weather. On another day, that player may perform poorly because of an injury. Similarly, yearly stats also fluctuate. A quarterback may throw for 5,000 yards one year, and the next year they may throw for 3,000 yards. Due to this, it is important to have a reliable application that can aid users of betting and fantasy football in which players they should put bet on, or choose for their fantasy teams. My motivation behind this project was to create something that could further betting and fantasy football, and even increase traction. My approach to doing this was putting together dataframes of yearly statistics from 75 quarterbacks, and seeing if my linear regression model could predict statistics for a future season. Reliable sources such as the NFL website, or ESPN, were used in order to put these statistics together. The main objective was to produce as accurate as possible stats, which came over time as more seasons were added, and the code was tweaked.

Explore More!

Research Paper

Project Webpage