In an era where professional sports franchises are increasingly reliant on analytics to build championship rosters, a group of students at The Ohio State University has proven that the future of sports data science is in capable hands. Junyi Li, an econometrics major with a perfect 4.0 GPA, recently led a team of five to deliver an ambitious “NBA Winning Percentage Prediction System” – a project that demonstrates how rigorous academic training can translate into practical, high-impact analytical tools.
From Classroom to Court
The project, completed over a three-month period from July to September 2023, was no ordinary academic exercise. Li and his team set out to build a predictive model capable of forecasting NBA team performance using regression sequence analysis. The goal was ambitious: to move beyond basic statistics and develop a system that could uncover the underlying relationships between team metrics and game outcomes.
What sets this project apart is the technical depth and leadership involved. As the team leader, Li took full ownership of the project lifecycle – from conceptual design to final review. The team created an end-to-end Winning Prediction System, building the architecture from scratch and navigating the complexities of sports data modeling.
“Regression analysis and econometric modeling are core to my academic training,” Li explained. “The NBA project gave me the opportunity to apply these techniques in a real-world context, where data is messy, variables are interrelated, and the difference between a winning and losing prediction can come down to subtle statistical signals.”
Technical Implementation and Innovation
The technical stack employed by the team is noteworthy for its blend of traditional and modern tools. Using Python and R for the heavy lifting of statistical modeling and data processing, the team also utilized HTML to create a front-end interface that made their predictions accessible and interpretable. This demonstrates the team’s ability to deliver not just a mathematical model, but an actual product that end-users could interact with.
One of the key challenges the team faced was ensuring model interpretability. In predictive analytics, a “black box” model is often less valuable than one that can explain why it makes certain predictions. Li and his team “optimized critical modules for model accuracy, utilization and interpretability” – suggesting they spent significant effort in making sure their findings were not just accurate, but actionable.
Leadership and Review
Perhaps the most striking aspect of Li’s contribution is his leadership. In a group project, coordination and vision are often as important as technical skills. Li “championed the review process with adviser,” acting as the primary liaison with faculty oversight and ensuring that the project met rigorous academic and practical standards.
The project was designed from the ground up, meaning Li and his team had to make fundamental decisions about data sources, variable selection, and modeling techniques without relying on existing templates. This “from scratch” approach is particularly challenging but also forces students to truly understand every component of their system.
Implications for the Future of Sports Analytics
This project arrives at a time when the intersection of data science and sports has never been more prominent. From the NBA’s adoption of player-tracking technology to the rise of advanced metrics like Player Efficiency Rating (PER) and Real Plus-Minus (RPM), teams are investing heavily in analytics departments. Predictive models like the one developed by Li and his team offer a glimpse into how machine learning and regression techniques could be used to inform strategic decisions – from lineup optimization to in-game tactical adjustments.
The Profile of a Data-Driven Leader
For Junyi Li, this project is part of a broader trajectory that includes hands-on experience in the financial sector. His internships at Hyde Park Investment Services Inc. and Insight Capital have honed his skills in statistical modeling and data-driven decision-making – skills that are directly transferable to the sports analytics world.
Li’s ability to lead a team, build a system from scratch, and communicate complex findings to both technical and non-technical stakeholders marks him as a rising talent in the field of quantitative analysis. Whether his career leads him to Wall Street or the front office of an NBA franchise, it is clear that projects like these are building the foundation for a data-driven future.
