Johns Hopkins University’s SARG Advances Sports Data Analytics Through Advanced Optimization and Computer Vision

Researchers and students at Johns Hopkins University’s Sports Analytics Research Group (SARG) are leveraging advanced combinatorial mathematics and computer vision to transform professional sports operations. By developing tools like the Lineup Optimizer and automated equipment measurement systems, the group demonstrates the practical application of data-to-wisdom pipelines in high-stakes environments. This work highlights the growing reliance on quantitative analysis in the Data & Analytics sector to drive decision-making in everything from field strategy to logistics.

Led by Anton Dahbura, SARG has developed the Lineup Optimizer, a tool that utilizes combinatorial mathematics to determine the most effective batting sequences for baseball teams. By modeling hundreds of thousands of permutations based on individual player statistics, the software challenges traditional coaching instincts, such as suggesting that sluggers might be more effective in the lead-off spot than traditional fast runners. This project exemplifies the shift from raw data to actionable knowledge, providing a framework that can be applied across various levels of play, from high school to the major leagues.

Beyond on-field strategy, the group has applied high-performance computing to complex logistical challenges like league scheduling. Originally focused on the minor leagues, the Baseball Scheduling Optimization group, co-led by Associate Research Professor Donniell Fishkind, uses supercomputers to generate optimized schedules in minutes—a task that previously took human experts weeks. While Major League Baseball took over minor league scheduling in 2020, SARG continues to provide these data-driven services to independent leagues and national junior hockey organizations.

SARG’s collaboration with professional organizations like the Baltimore Orioles has resulted in specialized tools such as an automated bat measurement system. Developed by students Kevin Wu and Jason Sun, the system uses computer vision to measure bat dimensions with 99.8% accuracy in seconds, replacing error-prone manual processes that previously took up to an hour. Sig Mejdal, who oversees the Orioles' analytics department, noted that such projects provide the team with skilled analysts while offering students real-world experience. The program currently supports over 20 projects across multiple sports, including football, soccer, and Formula One racing, serving as a talent pipeline for the broader analytics industry.