Kabir Devgun

Pro Football Focus is a sports analytics company that aggregates and performs advanced analytics on game-day data from the NFL and NCAA Division One football. I have been invited to join their academic research cohort, where I am currently conducting self-guided research. My main research hypothesis centers around evaluating how good a player is in coverage. I am leveraging play-by-play, quarterback charting, and all-coverage PFF-sponsored dataset to design a more stable and accurate measure of a player's coverage production, independent of the result of the play.

I interned with the team's research and development group, where I was accountable for building a model that predicted the expected number of games likely to be missed by a player due to injury based on a sample of historical injury data. My responsibilities included data engineering tasks, such as sanitizing and cleaning up stray data points and outliers, using various Python packages. I created and operationalized a model to project the expected number of games missed based on the specifics of that player's injury. This model is currently being leveraged by the team to evaluate a player's injury risk going forward and make decisions on contract status.

Titletown Tech is a venture fund created by a partnership between the Green Bay Packers of the NFL and Microsoft Corporation. I had an opportunity to work with one of their portfolio companies, Illative, where I developed a metric to rank NFL cornerbacks based on how closely they were able to track their opponents. The trackability metric was calculated from the player positions on the field for all plays overall in a full NFL season. This position tracking data was obtained from NFL NextGen Stats. My ranking metric correlated very well with established cornerback ranks, showcasing a good validation of my technique, as well as providing new data-driven input to establish rankings based on NextGen Stats.

I interned with the database management startup 601 Analytics, where I was tasked with managing data from events at Kaseya center (games, concerts, etc.) Using SQL, and generating reports using Power BI. I also worked on a pitch centered around Generative AI, and how it could be used to improve business operations and increase fan engagement.

EXL is an analytics consulting company. I was assigned to a team that was part of a large fan engagement program for the National Basketball League (NBA). My responsibilities included building a statistical model that predicted the probability of a fan purchasing an NBA League Pass subscription based on various variables. Tasks completed included creating a data frame containing the distance between every U.S. Zip code and every NBA arena in the U.S., and then sorting NBA fan activity data (variables included: NBA Store purchases, Gender, NBA All-Star Voting, etc.), and creating different types of distribution tables for visualization.