PCOSmart

Inspiration

For this women-centric hackathon, we wanted to focus on a niche, real-world issue affecting primarly women and were inspired to create PCOSmart. The cause of PCOS is unknown, and there is no cure; only symptom management. To help individuals easily assess their risk for Polycystic Ovary Syndrome (PCOS), a condition affecting millions, we developed a platform where users can easily seek guidance on how to move forward in a healthy and financially-feasible way. Our goal is to provide instant insights and personalized advice, making health information more accessible.

What it does

PCOSmart uses AI to predict the likelihood of PCOS based on user-reported symptoms and demographics. It provides personalized health and lifestyle advice also based off of input, empowering users to make informed decisions about managing their condition.

Challenges

We ran into challenges integrating OpenAI with neural network-based prediction model. Issues arose with processing user inputs and ensuring secure handling of sensitive data, like the OpenAI API key. We also were having issues with React, so we moved to using Next.js instead.

Accomplishments

We achieved an intuitive interface that has the potential to allow users to easily interact with the platform, as well as integrating a FNN model of 89% accuracy with OpenAI!

What we learned

We gained experience in AI integration, integrating frontend and backend code, and the importance of clear, empathetic communication, especially when dealing with health-related advice.

What's next for PCOSmart

If we had more time, integrating all of our parts together (we couldn't due to time constraints) would be the first step. Later in the future, we would like to create a multimodel AI via ultrasound imaging data and incorporate that into our risk assessment generation. Also, we would like to run the output of our model by real doctors to preserve health information accuracy.