-
-
The app predicts diabetes risk based on user input using a machine learning model.
Accomplishments that we're proud of
Inspiration
Many people are unaware of their risk for diabetes until symptoms appear. Early detection can lead to better treatment and lifestyle changes. We wanted to build a simple, AI-powered tool to help users assess their diabetes risk using common health metrics.
What it does
The app takes user input such as age, glucose levels, BMI, and other health-related features, and predicts whether the user is at risk of diabetes using a trained Random Forest classifier.
How we built it
We used Python, Pandas, Scikit-learn, and Streamlit to build the machine learning model and the web interface. The model was trained on the publicly available PIMA Indian Diabetes Dataset. We deployed the app using Streamlit Cloud and integrated it with Google Cloud for scalability.
Challenges we ran into
One challenge was cleaning the dataset and handling zero values in health metrics like insulin and skin thickness. We also had to tune our model to reduce false positives. Learning how to deploy the app smoothly using Streamlit and GCP took time but was worth it.
Accomplishments that we’re proud of
We created a clean and responsive app that works well for non-technical users and gives quick predictions. It was exciting to see everything come together—from data preprocessing to model training and final deployment.
What we learned
We learned how to apply machine learning in a health-related context, use Google Cloud tools for deployment, and improve collaboration through Git and Devpost.
What’s next
We plan to:
- Add more features like heart disease risk
- Improve model accuracy
- Store user data anonymously (if permitted) for future insights
Log in or sign up for Devpost to join the conversation.