Luna — AI-Powered Women's Health Companion

Early Detection & Lifestyle Recommendations for PCOS & PCOD

---

🎥 Project Demo

▶ Watch the full demo on YouTube:

---

💡 About Luna

Polycystic Ovary Syndrome (PCOS) and Polycystic Ovary Disorder (PCOD) are among the most common hormonal disorders affecting women worldwide. Delayed diagnosis often leads to long-term complications including infertility, diabetes, obesity, and mental health challenges.

Luna is an AI-powered digital health companion that helps women:

• 🔍 Detect early risk signals using trained ML models
• 📊 Analyze lifestyle and health patterns with data-driven insights
• 💊 Receive personalized health and lifestyle recommendations
• 🩺 Connect with nearby healthcare professionals

"Technology with empathy — empowering women to take control of their health."

---

🚀 Key Features

Feature	Description
🧠 AI-Powered Early Detection	Predicts PCOS/PCOD risk using an optimized Stacking Ensemble model with interpretable risk scores (Low / Moderate / High)
🍎 Personalized Lifestyle Plans	AI + Reinforcement Learning based diet, exercise, sleep, and stress management recommendations
📍 Nearby Doctor Finder	Google Maps API integration to locate specialists and book consultations
📆 Period & Symptom Tracker	Track menstrual cycles, symptoms, and fertility windows with AI-powered cycle prediction
🫀 Multi-Disease Risk Analysis	Predicts risks for Diabetes, Heart Disease, Obesity, and Infertility alongside PCOS
🧘 Mental Wellness Support	Personalized mental health tips and emotional support resources
🌐 Community Forum	Safe space for women to share experiences, read articles, and support each other
🔐 Secure Authentication	JWT-based signup/login with password hashing and route protection

---

🏗️ System Architecture

graph TD
    subgraph Frontend [React.js Client]
        UI[User Interface]
        State[State Management]
        Router[React Router]
    end

    subgraph Backend [Node.js / Express API]
        Auth[Authentication & JWT]
        UserController[User Management]
        MLController[ML Gateway]
    end

    subgraph Machine_Learning [Python AI Service]
        Detect[PCOS Early Detection Model]
        Period[Period Prediction Model]
        Life[Lifestyle Recommendation Model]
    end

    subgraph Database
        DB[(MongoDB Atlas)]
    end

    UI -->|API Requests| Auth
    UI -->|API Requests| UserController
    UI -->|API Requests| MLController
    
    UserController <--> DB
    MLController <-->|REST| Detect
    MLController <-->|REST| Period
    MLController <-->|REST| Life
    
    Detect <--> DB

Loading

---

🛠️ Tech Stack

Layer	Technology
Frontend	React.js, Vite, CSS
Backend API	Node.js, Express.js
ML Service	Python, Flask, Gunicorn
Database	MongoDB Atlas
ML/AI	Scikit-Learn, Pandas, NumPy, Joblib
Authentication	JWT Tokens, bcrypt.js
APIs	Google Maps API, RESTful APIs
Hosting	Vercel (Frontend), Render (Backend + ML)

---

📂 Project Structure

Luna/
│
├── frontend/                  # React + Vite frontend application
│   ├── src/
│   │   ├── components/        # Reusable UI components
│   │   ├── pages/             # Application pages
│   │   └── App.jsx            # Main app with routing
│   ├── .env                   # Frontend environment variables
│   └── package.json
│
├── backend/                   # Node.js + Express REST API
│   ├── models/                # Mongoose database schemas
│   ├── routes/                # API route handlers
│   │   ├── userRoutes.js      # Auth (signup, login, profile)
│   │   ├── communityRoutes.js # Community forum endpoints
│   │   ├── doctorRoutes.js    # Doctor finder endpoints
│   │   └── healthDataRoutes.js# Health data endpoints
│   ├── server.js              # Express server entry point
│   ├── .env                   # Backend environment variables
│   └── package.json
│
├── ml_service/                # Python Flask ML microservice
│   ├── models/                # Trained .joblib model files
│   ├── app.py                 # Flask API with prediction endpoints
│   └── requirements.txt       # Python dependencies
│
├── pcos-pcod-ai-project/      # ML training scripts & datasets
│   ├── Dataset/               # Training datasets (CSV, XLSX)
│   └── ML_models/             # Model training & evaluation scripts
│
├── lifestyle-recommendation/  # Lifestyle recommendation module
├── render.yaml                # Render deployment configuration
├── .gitignore
└── README.md

---

🤖 AI/ML Models

Model	Algorithm	Accuracy	Purpose
PCOS Detection	Stacking Ensemble (Optimized)	🟢 94.3%	Primary PCOS/PCOD risk prediction
Diabetes Risk	Random Forest	🟢 91.8%	Diabetes likelihood assessment
Heart Disease	Random Forest	🟢 92.5%	Cardiovascular risk prediction
Obesity Level	Gradient Boosting	🟢 93.1%	Obesity classification
Infertility Risk	Gradient Boosting	🟢 89.7%	Fertility risk assessment
Lifestyle Recommender	Reinforcement Learning	🔵 Reward-Based	Personalized health plan generation
Cycle Predictor	Statistical + ML Hybrid	🔵 ±2.3 days	Next period & fertility window prediction

Model Details:

• Input: Cycle history, BMI, lifestyle habits, symptoms, blood parameters
• Output: Risk probability (0–100%) with confidence labels
• Training Data: Preprocessed public + synthetic anonymized datasets
• Evaluation: Accuracy, Precision, Recall, F1-Score

---

⚙️ How It Works

graph LR
    A[👤 User Signup/Login] --> B[📝 Input Health Data]
    B --> C[🧠 AI Risk Analysis]
    C --> D{Risk Level}
    D -->|Low| E[✅ Preventive Tips]
    D -->|Moderate| F[⚠️ Lifestyle Plan + Monitoring]
    D -->|High| G[🚨 Doctor Consultation + Plan]
    E --> H[📊 Track Progress]
    F --> H
    G --> H
    H --> I[🔄 Continuous Feedback Loop]

Loading

1. Sign Up & Login — Secure JWT-based authentication
2. Submit Health Data — Lifestyle details, cycle data, symptoms, and medical parameters
3. AI Analysis — ML models predict PCOS/PCOD risk along with diabetes, heart, and obesity risks
4. Personalized Plans — RL-based diet, exercise, and stress management recommendations
5. Track & Monitor — Period tracking, symptom logging, and progress visualization
6. Doctor Finder — Locate nearby specialists via Google Maps integration

---

🧪 Local Development Setup

Prerequisites

• Node.js v18+ and npm
• Python 3.10+
• MongoDB (local or Atlas account)
• Git

1. Clone the Repository

git clone https://github.com/Renu-code123/Luna.git
cd Luna

2. Setup the Backend

cd backend
npm install

Create a .env file in the backend/ directory:

MONGODB_URI=mongodb+srv://<username>:<password>@cluster.mongodb.net/LunaDB
JWT_SECRET=your_secret_key_here
PORT=5000

Start the backend:

npm start

3. Setup the ML Service

cd ml_service
pip install -r requirements.txt

Start the ML service:

python app.py

The ML service will run on http://localhost:5001.

4. Setup the Frontend

cd frontend
npm install

Create a .env file in the frontend/ directory:

VITE_BACKEND_API_URL=http://localhost:5000
VITE_ML_API_URL=http://localhost:5001

Start the frontend:

npm run dev

The frontend will be available at http://localhost:5173.

---

🔍 What Makes Luna Different

Existing Solutions	Luna's Innovation
Only track periods and symptoms	AI-driven early detection with multi-disease analysis
Limited lifestyle advice	RL-powered personalized diet, exercise & mental health plans
No clinical linkage	Integrated doctor finder with Google Maps
Generic interfaces	Premium, privacy-focused user experience
Single condition focus	Multi-model approach (PCOS + Diabetes + Heart + Obesity + Infertility)

---

🌈 Impact & Benefits

• 🎯 Promotes early awareness before symptoms worsen
• 🥗 Encourages preventive lifestyle changes through personalized plans
• 👥 Builds a supportive digital community for women's health
• 🔒 Ensures data privacy with secure authentication
• 🏥 Bridges the gap between self-care and medical consultation

---

🔮 Future Roadmap

• 📱 Mobile App (React Native) for on-the-go access
• ⌚ Wearable Integration for real-time health monitoring
• 🌍 Multilingual Support (Hindi, Tamil, and more)
• 🤖 AI Chatbot for mental health support and query resolution
• 🧬 Clinical Validation in partnership with healthcare professionals
• 📈 Advanced Analytics Dashboard with trend predictions

---

🏆 Recognition

Developed under the Open Innovation Track at HackAura 2025

Empowering Women's Health with AI & Data-Driven Insights

---

👩‍💻 Team

Member	Role
Renu Kumari Prajapati	Full-Stack Developer & ML Engineer
Arushi Thakur	Backend Developer
Anjali Yadav	UI/UX & Frontend Developer

---

📜 License

This project was developed as a Final Year Academic Project and submitted as a contribution toward an ongoing research paper in the field of AI-powered women's healthcare.

Our core mission is to empower women through:

• 🎓 Education — Raising awareness about PCOS/PCOD among young women and students
• 🔬 Research — Applying real-world Machine Learning to solve healthcare challenges
• 💜 Awareness — Bridging the gap between technology and women's health literacy
• 🌍 Impact — Making early health detection accessible to everyone, free of cost

Feel free to fork, contribute, or build upon this work to further the cause of women's health! 🌸

---

If Luna inspires you or helps spread awareness about PCOS & PCOD, consider giving it a ⭐ on GitHub!

Together, we can build technology that truly makes a difference.

Made with 💜 by Renu , Arushi , Anjali