Dartmouth and UMass Amherst students shine at 2024 SMT Data Challenge undergraduate division

From Saturday, November 16, to Tuesday, November 19, in Durham, North Carolina, three talented undergraduates—two from Dartmouth and one from UMass Amherst—participated in the SportsMEDIA Technology (SMT) DATA Challenge Finalist VIP Weekend. The team included Atul Venkatesh '27, a Quantitative Social Science (QSS) major and Head of Research at Dartmouth Sports Analytics; Levon Sarian '27, a Computer Science major from Dartmouth; and Ishan Kinikar '27, a Computer Science major studying at UMass Amherst.  

Ishan, a computer science major with an interest in sports, was invited by Atul to join a competition. These two have been close friends for over five years.  Atul met Lev at a pre-college event for members of the Class of 2027 in the Greater Boston area, and at this event they discovered their shared passion for sports and data. Since then, Atul and Lev have represented Dartmouth in a football hackathon hosted by Syracuse University and worked together on a part-time soccer analytics job. As someone looking to pursue a career in sports analytics, Atul seeks opportunities to showcase his skills through projects and competitions. When the SMT competition appeared on his LinkedIn feed, he recognized it as a great opportunity and reached out to Lev and Ishan to collaborate.

SMT hosted a month-long data challenge during the past summer, providing extensive datasets and encouraging students to create innovative projects on any topic in baseball. The competition featured two divisions. undergraduate and graduate. Participants were tasked with sharing their code on GitHub, writing a 2,000-word paper, and developing an interactive user interface for exploration. Projects were judged by a panel of experts, and out of approximately 75 teams, three from each division were invited to the Finalist VIP Weekend. During this final round, participating student presented their projects to a new panel of judges, with one winner from each division crowned based on a combination of presentation and written evaluation scores.  

Atul shared his experience of the conference in his own words: 

The purpose of the conference was for the finalists to present their projects, and the entire weekend was designed by SMT to celebrate the finalists. Conference finalists attended a Carolina Hurricanes game, toured the Durham Bull stadium, and toured SMT headquarters.

The climax of the weekend was on Monday night, where SMT rented out a restaurant, and we presented in front of the panel of judges. The judges' scores were averaged with written evaluation scores, and the team with the highest overall score was declared the winner. In the end, our team (Atul Venkatesh '27, Levon Sarian '27, and Ishan Kinikar UMass Amherst '27) were declared the winners.

The conference was a once-in-a-lifetime experience, and SMT did an excellent job of making us feel special. It was also a pleasure to meet many other competitors and see their final products. Everyone was super creative, and there was a great deal of diversity in terms of project ideas. We would recommend everyone interested in data science or sports analytics to participate in this competition.

We used play-by-play data to determine the right time to pull a pitcher based on situational factors such as pitch count, pitches in an inning, outs in an inning, baserunner situation, and the hitter due up. This was done by creating multiple XGBoost algorithms. Our ultimate product was an interactive user interface called "The Snell Tool,"where users can input a hypothetical situation, and our model will decide if the pitcher should be kept in or pulled out. We believe that with some refinement, our tool can be an important asset for managers in the heat of the game.

Check out The Snell Tool here.

For more details on the competition results, read the official press release from SMT.