Implementation of a recruit visualization tool for UVA football

摘要

Effective recruiting is essential for building a successful football program, as it is the primary method by which college football teams acquire new players. The University of Virginia football recruiting staff works hard to direct its efforts toward recruits that fit Virginia's football culture and can lead to future wins. This paper describes four models and visuals that help the staff determine which recruits to pursue. Recruiters and coaches evaluate players based on the following criteria: 1) football talent, 2) academic performance, and 3) mental toughness, referred to as “grit”. By analyzing the staff's processes and metrics, we found areas where data analytics could be effectively applied. A model that forecasts a recruit's college GPA was developed to help predict a recruit's future academic success. To inform grit evaluations, we developed a visual that uses recruits' tweets to show their confidence and toughness. Additionally, to ensure recruiting efforts are directed at players that have a chance of committing to Virginia, we designed a prediction model that outputs the likelihood of a recruit selecting the University of Virginia from the set of schools that have given the recruit an offer to join the team. Furthermore, a supplementary visual known as the competitive landscape plots Virginia against all other schools that have given the recruit an offer, helping the coaches see how they differ from competing schools. Ultimately, these models and visuals were combined into a comprehensive visual, called the Recruit Dashboard, to give the staff a simple way to analyze all metrics needed to evaluate a recruit.