🩺 Health Risk Prediction App

AI-powered health risk prediction and chatbot system

---

Important

📋 Version Updates from v1.0:

1. Interactive Frontend Design: Features a responsive layout with Tailwind CSS, a side-overlay chatbot, and smooth form interactions.
2. Health Risk Prediction: Integrates a RandomForestClassifier model trained on the PIMA Indians Diabetes Dataset, logged via MLflow.
3. RAG Chatbot: Implements a local RAG system using Ollama (Llama 3.2 1B and nomic-embed-text) for health-related queries.
4. CORS Middleware: Added FastAPI CORS support for seamless communication between frontend (http://localhost:5173) and backends (http://localhost:8000, http://localhost:8001).
5. CI/CD Integration: GitHub Actions automates model training, testing, and deployment for continuous improvement of ML models.

---

📚 Table of Contents

• Overview
• Technical Flow Chart
• Key Features
• Technology Stack
• Installation and Setup
  • Using Docker
  • Manual Installation
• Usage
• Contributions
• License
• Contact

---

📌 Overview

The Health Risk Prediction App is an AI-powered web application designed to predict health risks (e.g., diabetes) based on user-submitted health metrics and provide health-related information via a RAG-powered chatbot.

🚀 Powered by Local AI and ML, this system integrates:

• 🤖 Machine Learning: RandomForestClassifier for health risk prediction, tracked with MLflow.
• 📚 RAG Chatbot: Local Ollama models (Llama 3.2 1B and nomic-embed-text) for answering health queries using ingested documents.
• 🌐 Responsive Frontend: Built with React and Tailwind CSS, featuring a clean UI and side-overlay chatbot.
• ⚡ Scalable Backend: FastAPI servers for health data management and RAG query processing, with PostgreSQL for data storage.

Health Risk Prediction App - Screenshots

Below are two screenshots of the Health Risk Prediction App.

What You’ll Learn from This Project 📖

🔹 👨‍💻 Machine Learning Integration: Train and deploy ML models with MLflow and FastAPI.
🔹 🔍 RAG Implementation: Build a local RAG system for health data queries.
🔹 ⚡ Modern Frontend Design: Create responsive UIs with React and Tailwind CSS.
🔹 🔒 Scalable Backend: Use FastAPI and SQLAlchemy for robust data handling.

📂 For learners: Refer to health-app-backend/app/main.py, backend/api.py, and frontend/src/App.jsx for detailed code comments explaining the ML and RAG workflows!

---

🛡️ Technical Flow Chart

Flowchart: The system collects health data via a React frontend, sends it to a FastAPI backend for storage and ML-based risk prediction, and uses a separate FastAPI server with Ollama for RAG-based chatbot responses from ingested health documents.

---

✨ Key Features

• 🤖 Health Risk Prediction: Predicts diabetes risk using a RandomForestClassifier trained on the PIMA Indians Diabetes Dataset, logged via MLflow.
• 📚 RAG-Powered Chatbot: Answers health-related queries using local Ollama models and ingested documents (e.g., PDFs).
• 💻 Responsive UI: React frontend with Tailwind CSS, featuring a health data form, prediction display, and side-overlay chatbot.
• ⚡ Continuous Model Improvement: CI/CD pipeline with GitHub Actions automates model training and evaluation.
• 🔄 Data Management: Stores health metrics in PostgreSQL via FastAPI and SQLAlchemy.
• 🌐 CORS Support: Ensures seamless frontend-backend communication across ports.
• 📜 Local AI Processing: Runs all AI models (ML and RAG) locally for privacy.

Note

Upcoming features:

1. Support for additional health risk models (e.g., heart disease).
2. Enhanced RAG with larger document ingestion.
3. User authentication for personalized data tracking.
4. Open to suggestions and contributions.

---

🛠️ Technology Stack

Component	Technologies
🔹 Backend Framework	FastAPI, SQLAlchemy
🔹 Machine Learning	scikit-learn, MLflow
🔹 RAG System	Ollama (Llama 3.2 1B, nomic-embed-text), LangChain
🔹 Database	PostgreSQL
🔹 Frontend	React, Tailwind CSS, react-router-dom
🔹 Deployment	Vite (frontend), Uvicorn (backend), Docker

Backend Dependencies (from health-app-backend/requirements.txt):

• fastapi==0.115.0, uvicorn==0.30.6, sqlalchemy==2.0.35
• mlflow==2.16.2, scikit-learn==1.5.2, pandas==2.2.3
• ollama==0.4.8, langchain==0.3.3, psycopg2-binary==2.9.9
• Full list in health-app-backend/requirements.txt.

Frontend Dependencies (from frontend/package.json):

• react==18.2.0, react-dom==18.2.0
• axios==1.7.2, tailwindcss==3.4.13, react-router-dom==6.26.2
• Full list in frontend/package.json.

---

🚀 Installation & Setup

📌 Option 1: Using Docker

Prerequisites:

• Docker installed
• Ollama running separately for LLM support
• PostgreSQL instance (local or cloud)

1️⃣ Clone the Repository

git clone https://github.com/TyrelM10/healthApp.git
cd healthApp

2️⃣ Create Environment Files

Create health-app-backend/.env:

DATABASE_URL=postgresql://user:password@localhost:5432/health_db
MLFLOW_TRACKING_URI=http://localhost:5000

Create backend/.env:

OLLAMA_HOST=http://localhost:11434

3️⃣ Set Up Ollama

Run Ollama in a separate terminal or as a service:

ollama serve
ollama pull llama3.2:1b
ollama pull nomic-embed-text

Verify:

ollama list

Ensure llama3.2:1b and nomic-embed-text are listed.

4️⃣ Build the Docker Images

cd health-app-backend
docker build -t health-app-backend .
cd ../backend
docker build -t health-rag-backend .
cd ../frontend
docker build -t health-app-frontend .

5️⃣ Run the Docker Containers

docker run -d --name health-app-backend -p 8001:8000 --env-file health-app-backend/.env health-app-backend
docker run -d --name health-rag-backend -p 8000:8000 --env-file backend/.env health-rag-backend
docker run -d --name health-app-frontend -p 5173:5173 health-app-frontend

• Health Backend: http://localhost:8001
• RAG Backend: http://localhost:8000
• Frontend: http://localhost:5173

Managing the Containers:

Stop the Containers

docker stop health-app-backend health-rag-backend health-app-frontend

Start the Containers

docker start health-app-backend health-rag-backend health-app-frontend

View Logs

docker logs health-app-backend
docker logs health-rag-backend
docker logs health-app-frontend

Remove the Containers

docker rm health-app-backend health-rag-backend health-app-frontend

Troubleshooting:

Container Health Check

docker inspect --format='{{.State.Health.Status}}' health-app-backend

Container Not Starting

Check logs for errors:

docker logs health-app-backend

Ensure Ollama and PostgreSQL are running and environment variables are correct.

📌 Option 2: Manual Installation

Prerequisites:

• Python 3.9+
• Node.js 18+, npm 9+
• Ollama for local LLM support
• PostgreSQL instance
• MLflow server

1️⃣ Clone the Repository

git clone https://github.com/TyrelM10/healthApp.git
cd healthApp

2️⃣ Health Backend Setup

Create & Activate Virtual Environment

cd health-app-backend
python -m venv venv
source venv/bin/activate  # Mac/Linux
venv\Scripts\activate     # Windows

Install Dependencies

pip install -r requirements.txt

Set Up Environment Variables

Create health-app-backend/.env:

DATABASE_URL=postgresql://user:password@localhost:5432/health_db
MLFLOW_TRACKING_URI=http://localhost:5000

Run MLflow Server

mlflow server --host 0.0.0.0 --port 5000

Train Model

python app/ml/train.py

Run the Backend

uvicorn app.main:app --port 8001 --reload

Health Backend available at: http://localhost:8001

3️⃣ RAG Backend Setup

Create & Activate Virtual Environment

cd ../backend
python -m venv venv
source venv/bin/activate  # Mac/Linux
venv\Scripts\activate     # Windows

Install Dependencies

pip install -r requirements.txt

Set Up Environment Variables

Create backend/.env:

OLLAMA_HOST=http://localhost:11434

Run the Backend

uvicorn api:app --port 8000 --reload

RAG Backend available at: http://localhost:8000

4️⃣ Frontend Setup

Navigate to Frontend

cd ../frontend

Install Dependencies

npm install

Run the Frontend

npm run dev

Frontend available at: http://localhost:5173

5️⃣ Ollama Setup

Install Ollama

Follow instructions at ollama.ai.

Run Ollama Server

ollama serve

Pull Models

ollama pull llama3.2:1b
ollama pull nomic-embed-text

Verify

ollama list

Ensure llama3.2:1b and nomic-embed-text are listed.

---

🧠 Usage

Note

1. The first run may take time as Ollama downloads models.
2. Ensure Ollama, MLflow, and both backends home are running before starting the frontend.
3. Check browser console (F12) for frontend errors and backend terminals for warnings.

• Open http://localhost:5173 to access the app.
• Navigate to Home for health tips.
• Use Health Data to submit metrics (age, BMI, glucose, etc.).
• Visit Predictions to get risk predictions.
• Interact with the Chatbot to ask health-related questions.
• Ingest health PDFs via http://localhost:8000/ingest or streamlit run backend/ingest_ui.py.

---

🤝 Contributions

Contributions are welcome! Check the issues tab for feature requests and improvements.

To contribute:

1. Fork the repository.
2. Create a feature branch (git checkout -b feature/YourFeature).
3. Commit changes (git commit -m 'Add YourFeature').
4. Push to the branch (git push origin feature/YourFeature).
5. Open a Pull Request.

---

⚖️ License

This project is licensed under the MIT License. See the LICENSE file for details.

---

📬 Contact

For questions or collaboration inquiries, reach out to Tyrel Menezes on:

🔗 LinkedIn: https://www.linkedin.com/in/tyrel-menezes
🔗 GitHub: https://github.com/TyrelM10

🔝 Return