Data that Matters: Using Computer Vision to Measure Engagement and Participation in Youth Sports


The statistics on childhood physical activity show that 80% of young people in the US are classified as “inactive”. More than 70% of kids drop out of youth sports by age 11, with Girls being 50% more likely to drop out than Boys. While organized sports provide a source of physical activity for some kids, not everyone has access to those types of programs due to their cost. The number one reason young people who stop a sport give is that “it’s no longer fun.” We used Machine Learning to create a data analytics tool that can provide parents and amateur coaches the possibility to measure the level of a child’s engagement with the sport. We process large amount of video data that have been collected, labeled and processed by experts in the domains of competence (i.e. boys/girls youth basketball and soccer). Our models use Computation Neural Networks to extract actionable predictions for coaches and parents in relation to a kid’s engagement with the sport. Initial test results show precision and accuracy in classifying input videos into labeled categories and measuring an “Engagement” index for each kid over time. Current and future work will involve Body Pose Estimation to measure engagement with sports on a new level of features.


Giovanni Liotta
CEO, Product Development, Dojo Science, Missouri, United States


Presentation Type

Innovation Showcase


Gender Equity and Policy


Youth Sports, Machine Learning, Engagement

Digital Media

This presenter hasn’t added media.
Request media and follow this presentation.