Nova | Devpost

Company logo

🚀 Inspiration Every 23 seconds, someone in the U.S. is diagnosed with diabetes. What if we could predict and prevent it before it happens? Type 2 diabetes, which accounts for about 90–95% of all diabetes cases, is largely preventable through lifestyle changes and can be reduced in up to 58% of high-risk individuals. Considering that many health-tracking apps focus on management rather than prevention, we were inspired to do the opposite.

🤖 What It Does Nova is a personalized AI medical assistant designed to help prevent lifestyle diseases like diabetes. Users input their daily habits—such as food intake and exercise—and Nova analyzes this data to calculate a real-time diabetes risk score (0–100). It also tracks key health metrics (steps, calories, glucose) and allows direct communication with healthcare providers for early intervention.

🛠️ How We Built It The Nova web-based app was developed using HTML and CSS for the frontend, creating a clean and intuitive user interface. We utilized React for building dynamic components and Node.js for the framework and deployment, ensuring smooth performance and scalability. JavaScript played a key role on both the frontend and backend, allowing seamless data flow and interaction between components.

For version control and collaboration, we used Git and managed our codebase efficiently with VS Code. To power Nova’s AI capabilities, we integrated OpenAI’s API. This combination of technologies enabled us to create an interactive, data-driven, and user-friendly health assistant.

🐞 Challenges We Ran Into Creating a robust application like Nova posed several challenges, including: Ensuring the originality of our application in comparison to multiple other health-tracking systems Deploying the backend and server of the project to create full logic functionality.

🏆 Accomplishments That We're Proud Of

Prototype: Developing a working prototype of the AI chatbot that demonstrates calculating key health metrics, predicting diabetes risk, and visualizing the results.

Machine Learning: Using an existing diabetes dataset with variables such as glucose levels and age to train a logistic regression model that predicts diabetes risk. This model mimics the logic behind Nova’s AI-driven risk score calculation.

User-Friendly Interface: Designing an intuitive and visually appealing user interface that enhances the overall user experience.

📚 What We Learned Through this project, we deepened our understanding of machine learning algorithms, API integrations, and full-stack development. We also learned the importance of effective team collaboration when troubleshooting complex bugs. This experience strengthened our skills in problem-solving, health tech innovation, and building user-centric applications.

🌟 What’s Next for Nova Adding a feature to contact healthcare providers directly

Using AI to create personalized diet and workout plans based on health data

Ensuring security with salted hashing, multi-factor authentication (MFA), and data encryption (AES-256 and TLS 1.3) for HIPAA compliance

Controlling access with role-based access control (RBAC) and attribute-based access control (ABAC), along with activity tracking through audit logs

Expanding Nova to track additional lifestyle diseases such as heart disease, high blood pressure, and cancer