In exciting news, the San Francisco 49ers have hired recent Biostats graduate Matt Ploenzke, PhD’20, to their analytics team. Ploenzke, who’s previous work with Professor Rafa Irizarry focused on applying machine learning methods to cancer genomics, applied his analytical skills to tackle a new venture, namely winning the competitive Big Data Bowl. The Bowl is an annual sports analytics contest in which members of the analytics community are challenged to contribute to the NFL’s continuing evolution in the use of advanced analytics.
Using Next Gen Stats data to build interpretable model inputs based upon football-specific domain knowledge, Ploenke’s innovative approach for predicting run success was to consider global associations with yards gained rather than player-specific features.
The rationale for Ploenzke’s approach borrowed from a common healthcare situation where ‘black-box’ models used to determine a patient’s best treatment option ultimately rely upon the expertise of a medical professional to determine best course of action. According to Ploenzke, “the doctor’s ability to do this effectively stems from their experience and the process of doing so may be referred to as meaningful feature engineering; that is to say the doctor has learned relevant indicators of patient success.”
According to Ploenzke, “combining domain knowledge through meaningful feature engineering with the predictive performance of machine learning models is often the formula for a winning gameplan.” In the case of the Big Data Bowl, Ploenzke’s winning analysis highlighted the importance of ball carrier downfield acceleration and unblocked tackler distance and spacing as key factors for predicting run success. All of which goes to show, a solid foundation in statistical analytical methods can carry you far! Congratulations Matt!