Computer Science Team Scores in March Madness Data Analytics Challenge
March Madness got off to a roaring start as four Rose-Hulman computer science students applied their problem-solving and data analytical skills to earn honors in the NCAA’s Crossroads Classic Analytics Challenge that simulated the Final Four basketball experience.
Seniors William Gardner, Xander Good, Nathan Loafman, and Xianshun Jiang were members of the Rose Miners team that placed third within the undergraduate division of a competition among 67 teams from seven Indiana colleges and universities.
They were asked to create predictive models that could help NCAA officials identify potential ticket purchasers for games in all rounds of the 2024 Division I Women’s Basketball Championship, including the Final Four round, after examining customer datasets from primary or secondary markets of past tournament sites. Teams were encouraged to integrate external data to improve the predictive power of their models.
“We were ready for this challenge because of the many types of assignments we’re asked to tackle in our (computer science) classes. However, we were pleasantly surprised with how well we did. There was some stiff competition,” said Gardner, the team’s leader who has accepted a post-graduate job using his computing skills within the financial services area.
Loafman added, “This was really a real-world situation that pushed us to learn and apply some new software. That just required a little more work for us.” He is planning to work with Artisan Electronics after graduating this spring.
Jiang’s second major in data science was especially helpful in creating the analytical models that helped the team quality for the competition’s final-round presentations March 8 at NCAA Headquarters in downtown Indianapolis.
Each of the final qualifying teams practiced in separate conference rooms before making their final 10-minute presentation, along with a five-minute question-and-answer period, before a panel of judges of professionals from Salesforce Inc. and the NCAA. The scoring criteria, which wasn’t revealed until after the competition, featured (in order of importance) business application, creativity, presentation, technical skills, and visualization. The competition was co-sponsored by NCAA, Tableau, and Visualize Your Technology.
“Our team held their heads high and did a great job under some challenging circumstances,” said team advisor Olga Scrivner, PhD, assistant professor of computer science and software engineering. She was one of the competition’s faculty champions. “The raw data provided was very challenging, but our team was able to use data wrangling skills from our Data Mining course. They also had to learn many new skills, presenting data with Tableau software, learning about sports analytics, and writing an executive summary for business application,” she remarked.
Another faculty advisor who helped the team was John McSweeney, PhD, associate professor of mathematics.
A major challenge for the Rose-Hulman team was that the month-long competition took place at a time in the school calendar when students were concentrating on completing senior-year capstone projects, taking final exams for institute’s winter academic quarter, and finding time to relax before starting the spring term.
“We tried to squeeze in maybe an hour a day to work on our (NCAA challenge) project. We tried to complete as much as we could whenever we could under the circumstances,” stated Good.
The Rose Miners team will be sharing their model, presentation, and business solutions with members of the NCAA analytics unit.
This was the first year that Rose-Hulman has participated in the Crossroads Classic Analytics Challenge. The competition plans to expand to include teams from colleges and universities across the country, starting next year.
