Implementing data analytics for U.Va. Football

摘要

Advanced statistical models and data analytic tools to enhance football performance have not been utilized to their full potential. The University of Virginia Football program has begun to build analytics into their program; however, we believe that robust data analytics can continue to be added to further improve the decision-making process. This effort is supported by the belief of the coaching staff in analytic-based strategies. This paper outlines tools that we have built to enhance data analytics for U.Va. football. One of the tools outlined was built with the goal of removing the guess work out of play calling and providing the coaches with specific probabilities of potential outcomes of each play they could run. These probabilities are provided by statistical models, and a simulation tool has been built for the use by coaches and players. Two of the tools outlined will provide the coaches with analytics that enhance their scouting of opponents. One of these tools has been built to ensure that the coaches can quickly create powerful visualizations of their opponents play calling, while the other is to provide the coaches with critical statistics that they should focus on when game planning for an opponent. This paper also outlines how we have advised the coaches on their current data collection and how they can improve the way that they record data to suit their analytical needs. Providing the coaches with analysis that will improve on field performance requires providing them with intuitive tools that can be used without a data scientist present, while still ensuring that they understand how the analysis is conducted.