Over 180 student teams filled the Campus Instructional Facility at the University of Illinois Urbana-Champaign on March 28–30 to compete in the 2025 Illinois Statistics Datathon, one of the most anticipated data science competitions on campus. This year’s event, hosted in partnership with Synchrony, challenged students to develop predictive frameworks that could forecast customer spending while balancing the risks of credit overextension, fraud, and potential defaults.
The problem statement challenged students to analyze account data and behavioral attributes to address three key objectives: predicting customer spending for Q4 2025 based on eight months of recent data, classifying accounts into risk-based segments, and identifying potential credit line adjustments informed by predicted behavior and risk profiles.
“At our recent Datathon, we harnessed the power of tangible data and academic expertise to spark meaningful innovation,” said Deeptil Potluri, Vice President, Senior Solutions Architect Credit Technology Leader at Synchrony, who served as both a mentor and co-designer of the Datathon problem statement. “By bringing together diverse perspectives and fostering a collaborative environment, we are not only driving forward-looking research, but also laying a robust foundation for future generations. We are very glad to be a part of this effort.”
It was a challenging, real-world scenario with high stakes that students responded to with energy, determination, and technical creativity.
“This was hands down one of the most exciting campus events for the Synchrony team this year!” said Karin Dor Markovich, Synchrony’s AVP of Technology Centers and Early Career Development. “With our largest group of students signing up to participate, and all of the students being super engaged and asking really great questions during the actual event… The energy at CIF on Friday with approximately 650 participants was nothing short of amazing!”
2025 Datathon Winners
After a weekend of intense modeling, mentoring, and problem-solving, the final results were announced during a virtual award ceremony hosted by Synchrony on April 3. The top honors went to Ziyu Li (junior, mathematics), Zhanyang Jin (junior, computer science + mathematics), Tianqi Zhang (junior, statistics), and Chuoyue Li (sophomore, actuarial science and risk management) of team Canes.
Coming in second place this year was ATLAT with Jun Lee (sophomore, computer science+chemistry), Junhyun Back (sophomore, computer science), Seungmin Cho (senior, information sciences + data science), and James Lee (senior, information sciences + data science). In third place was STAT 510’s Best Students, James Xu (masters, statistics), Yanlin Wang (masters, statistics), Nay Petrucelli (masters, statistics), and Mehmet Korkmaz (masters, statistics).
Several other teams earned honorable mentions, and all participants were commended for their commitment and creativity.
Inside the Winning Strategy: Team Canes
Comprised of students with minimal prior exposure to financial data, team Canes approached the problem with humility and a strong focus on clarity and interpretability.
“The problems and data were complex since our team has little financial, economic, or business background,” the team shared. “We spent a lot of time trying to understand the data mapping document… this process lasted almost six hours, and we thought this was the most challenging part of the weekend.”
Rather than diving immediately into complex models, the team emphasized simplicity and interpretability — a strategy that ultimately paid off. Their approach broke the problem into three parts. First, they focused on forecasting Q4 spending using traditional data science pipelines that included feature engineering, model training, validation, and result integration. Second, they approached account classification as a clustering problem, enhancing their methods with neural networks for more effective feature creation. Finally, for credit line adjustment recommendations, they adopted a hybrid strategy that combined rule-based reasoning with machine learning techniques to balance accuracy with clarity.
The judges praised Canes for their outstanding and thorough analysis, highlighting the team’s exceptional technical execution and clear, well-structured presentation. They were particularly impressed by the team’s thoughtful evaluation of multiple models and the detailed rationale provided for their choices, such as selecting Random Forest for Q4 forecasting. Canes demonstrated a strong commitment to accuracy by validating their results and benchmarking performance from multiple angles. Their presentation not only set the context effectively but also clearly explained key decision points and proposed thoughtful next steps for further improvement.
“This achievement means a lot to us,” the team noted. “Beyond academic training, this is an excellent opportunity to be exposed to real-life data and industry problems. We appreciate it very much.”
A Partnership That Powers Opportunity
The ongoing collaboration between the Illinois Department of Statistics and Synchrony for the annual Datathon has grown into a cornerstone experience for many students. The Datathon is an opportunity for students to develop real solutions to complex problems using real data outside of the classroom. The event is also one that opens doors to internships and full-time careers.
“The Statistics Department really understands our company’s needs and how to create the best experience for the students,” said Dor Markovich. “They’ve done a great job throughout the years—refining our problem statements, offering workshops, and providing mentorship throughout the event.”
This partnership also feeds directly into Synchrony’s early career pipeline.
“We’ve seen a growth in applications from students who participated in the Datathon,” she added. “We’ve hired some as interns, and others have even landed full-time roles with us!”
For students interested in connecting with Synchrony, Dor Markovich encourages them to reach out at campus events, through email at etc@syf.com, or to connect with her directly on LinkedIn. “We have opportunities for students to intern with us all year round,” she said. “And just last week, we were recognized as #2 on Fortune’s 2025 list of 100 Best Companies to Work For in the U.S. I would love for students to learn more about who we are and our amazing culture.”
See the Datathon Landing Page for a full recap of Datathon winners, finalists, and honorable mentions, including links to their presentations.
Aaron Thompson is a staff writer for the Department of Statistics. If you have news to share, please contact the Statistics News Group at stat-marcom@illinois.edu