A pair of biomedical engineering students has leveraged their experiences in problem-based learning courses and skills from their computer science minors to win the 2021 CarMax Analytics Showcase.
Suraj Rajendran and Prathic Sundararajan used machine learning models to propose roughly 15 marketing and inventory strategies for CarMax, and they proposed a new technology platform the company could use to employ those models. They emerged from nearly 200 teams this month to win the $3,000 prize.
“Oftentimes, it’s difficult for students learning data science to get their hands on real-world challenges that are derived from industry-collected data,” said Rajendran, who’s in his final semester of studies in the Wallace H. Coulter Department of Biomedical Engineering. “This is especially true in the biomedical field and anything healthcare related due to privacy concerns. This was a great opportunity to practice similar skill sets in a slightly different industry.”
Rajendran and Sundararajan also worked with Benjamin John, a computer science student from the University of Illinois at Urbana-Champaign. They credited the Coulter Department’s problem-based learning approach for helping them tackle a very different kind of problem than they’ve encountered before.
“Early on, we used the methods taught through our BME classes and identified the needs of our stakeholders. This really helped put the whole project in perspective in terms of what we hoped to accomplish,” Sundararajan said.
“We identified novel methods we could apply to the given dataset to differentiate ourselves from the other teams and provide unique insights,” he said.
The trio incorporated publicly available information with the proprietary data CarMax provided to the teams. They also added the element of proposing a new software platform powered by their machine learning models.
“Going into the competition, our goal was to apply the practical skills we have learned in a real-world setting. Machine learning and data analytics are skills that are increasingly being used in the medical industry. Use of these techniques will not only allow for better data visualization, but also the creation of more computationally efficient models and devices,” Rajendran said.
The CarMax Analytics Showcase asked teams to use the company’s wealth of data to find unique insights and clearly present their results in answer to this question: “Given historical industry sales, how can CarMax tailor it’s marketing and inventory strategies to draw in distinct segments of consumers?” The top three teams presented to CarMax leaders, who selected the winner.
“We would definitely like to show appreciation to the biomedical engineering department at Tech. The project-based classes for the program have created an amazing environment for students and really challenge us to constantly keep learning new skills,” Sundararajan said. “Of course, this is only possible because of the amazing professors and academic advising staff — like Paul Fincannon and Brenda Morris — who really make the department as successful and well known as it is.”