Inspiration
Let’s face it, perfecting your basketball shot is hard. Whether you're just starting out or grinding solo sessions as a collegiate athlete, it's tough to know if your form is actually improving or getting worse. I personally struggled with this, never knowing if my form was solid or falling apart mid-shot. And with no coach around most of the time, I was stuck guessing. That’s why we built HoopHelper, an AI-driven shooting coach to give instant feedback and help anyone improve, from beginners to pros.
What it does
HoopHelper analyzes your shooting form in real-time using pose detection. It captures key body points, calculates shooting angles, and scores your shot based on ideal form metrics. You’ll receive personalized feedback after every shot, with tips on how to improve. You can also save your shots, track your progress, and review a wireframe visualization of your body alignment to see exactly what went wrong.
How we built it
We used React and Vite for the frontend to keep things fast and lightweight, paired with a Flask backend. For pose detection, we integrated YOLOv8s pose model, which lets us identify key body positions with high accuracy. This combination enables smooth, real-time analysis and feedback, making shooting practice smarter and more effective.
Challenges we ran into
One of our biggest challenges was time. With busy school schedules, we couldn't start working on the project until later in the week, which forced us to move quickly and prioritize efficiently. We also faced communication hurdles, so coordinating across our team was tough at times. It was also tough maintaining clear and consistent communication which slowed us down during key moments of development. Despite that, we pushed through and were able to bring HoopHelper to life.
Accomplishments that we're proud of
First and foremost, we’re incredibly proud to have included LeBron James, the King of Basketball and the undisputed GOAT, in our project (you’ll see). Beyond that, we’re stoked about how well the shooting form checker actually works. Even after just a few test runs, we noticed real improvement in our form. Seeing our idea help in a tangible way, even in such a short time, made all the late nights and debugging totally worth it.
What we learned
We gained valuable experience working with React and learning how to integrate multiple components across the frontend and backend. This project also gave us insight into project management principles with a longer timeline than a typical hackathon, we had to stay organized, plan ahead, and adapt when things didn’t go as expected. It was a great opportunity to practice collaboration, time management, and building something from idea to execution.
What's next for HoopHelper
Next, we’d like to add a layup analysis feature and introduce some gamification elements like badges, achievements, and progress tracking to keep users engaged and motivated. We also want to develop a parameter tuning system for coaches, allowing them to customize what “good form” looks like based on their training philosophy. While AI provides great feedback, we know there’s still an important human element in coaching.
Log in or sign up for Devpost to join the conversation.