Inspiration Diabetes is a growing global health concern, especially in developing countries like India. Early detection can help reduce risks and improve outcomes significantly. I wanted to build a simple, accessible tool to empower individuals to check their diabetes risk using machine learning — all wrapped in a beginner-friendly web app.
What it does This web app predicts whether a person is likely to have diabetes or not, based on inputs like glucose level, blood pressure, BMI, and more. It uses a trained ML model to give instant results in a user-friendly interface.
How we built it Frontend: Streamlit (Python framework) for UI.
Backend/Model: A machine learning model trained on the Pima Indians Diabetes dataset using scikit-learn.
Deployment: Deployed on Streamlit Cloud for easy access and sharing.
Challenges we ran into Finding and cleaning a quality dataset.
Making the Streamlit interface both responsive and user-friendly.
Ensuring the model wasn't overfitting and provided reliable predictions.
Accomplishments that we're proud of End-to-end implementation: from data preprocessing to ML model training and finally to deployment.
Deployed the app successfully online.
Built a project that could have real-life impact on health awareness.
What we learned Fundamentals of data preprocessing and model evaluation.
How to integrate ML with Streamlit for quick deployment.
Practical understanding of classification models like RandomForest.
What's next for DIABETES-PREDICTION-APP Improve model accuracy by using hyperparameter tuning and more advanced algorithms.
Add user login and data saving functionality.
Incorporate visualizations for better insights.
Expand to detect other diseases in future iterations.
Built With
- altair
- blinker
- certifi
- click
- colorama
- coloroma
- kagglehub
- matplotlib
- pandas
- pillow
- streamlit
- tenacity
Log in or sign up for Devpost to join the conversation.