Diabetes Risk Prediction App

• 

  The UI of the webapp

• 

  The UI with the prediction button clicked

Inspiration

Diabetes is a growing global health concern, affecting millions of people. Many individuals are unaware of their risk level until it's too late. Inspired by the need for early detection, we built this Diabetes Tester to provide a quick and accessible way to assess diabetes risk using machine learning.

What it does

This web-based application allows users to input key health metrics—such as glucose levels, BMI, and family history—to predict their risk of diabetes. The app: Accepts user health data via a simple form Uses a trained Random Forest model to make predictions Provides a diabetes risk score and classification (Diabetic or Non-Diabetic) Offers a user-friendly interface with real-time results

How we built it

Data Processing: Used a cleaned diabetes dataset with resampled data for better predictions

Machine Learning: Trained a Random Forest Classifier to classify patients based on input data

Framework: Developed the application using Streamlit for a simple yet effective user interface

Deployment: Hosted the app using GitHub + Streamlit Cloud

Challenges I ran into

Model Hosting Issues: Initially, Streamlit did not recognize joblib, requiring us to modify model loading logic. Missing Data Handling: Some users may not have Diabetes Pedigree Function, so we made it optional. GitHub File Paths: Directly loading .pkl files from GitHub required a workaround with urllib.request.

Accomplishments that we're proud of

Successfully built a fully functional and deployed diabetes risk predictor Improved model accuracy using balanced data Created a user-friendly UI that even non-technical users can navigate easily

What we learned

How to integrate machine learning models with Streamlit Best practices for handling missing data in real-world applications Deploying ML models on GitHub & Streamlit Cloud efficiently

What's next for Diabetes tester

Enhancing Model Accuracy by experimenting with different ML models like XGBoost Adding Data Visualization to provide more insights into diabetes risk factors Developing a Mobile App version for easier accessibility Multi-Language Support to make the app accessible globally

Built With

• joblib
• kaggle
• numpy
• pandas
• python
• scikit-learn
• streamlit