🩺 Cancer Risk Simulator

AI-Powered Cancer Risk Prediction with Real-Time ML Analysis

An interactive web application that predicts cancer risk using machine learning trained on real medical data with 93.7% accuracy

Features • Demo • Installation • Usage • Model

---

📋 Table of Contents

• Overview

• Features

• Demo

• Tech Stack

• Project Structure

• Installation

• Usage

• Model Details

• Disclaimer

---

🌟 Overview

The Cancer Risk Simulator is a full-stack machine learning application that provides real-time cancer risk predictions based on 8 key health factors. Built with a modern tech stack, it combines a powerful XGBoost model (trained on 1,500 real patient records) with an intuitive React interface featuring live visualizations and instant feedback.

🎯 Key Highlights

• 🧠 93.7% Accuracy - XGBoost model trained on real medical data
• ⚡ Real-Time Predictions - Instant risk calculation as you adjust parameters
• 📊 Interactive Visualizations - Beautiful charts with Recharts
• 🎨 Modern UI - Glassmorphism design with Tailwind CSS
• 🔄 Optimal Risk Comparison - See potential improvements with lifestyle changes
• 🚀 RESTful API - FastAPI backend with automatic documentation

---

✨ Features

🩺 Health Metrics Analysis

• 8 Comprehensive Factors: Age, Gender, BMI, Smoking Status, Genetic Risk, Physical Activity, Alcohol Intake, Cancer History
• Real-Time Updates: Predictions update automatically with debouncing (500ms)
• Risk Categorization: Low, Moderate, High, Very High with color-coded indicators

📈 Visualization & Insights

• Current vs Optimal Risk: Compare your risk with the best possible scenario
• Interactive Charts: Bar charts showing risk comparison
• Potential Improvement: Calculate possible risk reduction percentage
• Color-Coded Results: Visual feedback with dynamic color schemes

🔬 Machine Learning Backend

• XGBoost Classifier: Gradient boosting optimized for Mac M4
• ROC-AUC Score: 0.94+ on test data
• Feature Importance: Transparent model with interpretable features
• Automatic Model Loading: Serialized model (model.pkl) loaded on startup

🎨 User Experience

• Glassmorphism Design: Modern, elegant UI with blur effects
• Responsive Layout: Works seamlessly on desktop and mobile
• Smooth Animations: Polished transitions and hover effects
• Error Handling: Clear error messages and loading states

---

🎬 Demo

Live Interface

🖥️  Frontend: http://localhost:5173
📡  Backend API: http://localhost:8000
📖  API Docs: http://localhost:8000/docs

Example Prediction Flow

1. Adjust Health Metrics → 2. See Real-Time Risk Update → 3. View Optimal Comparison → 4. Analyze Improvement Potential

---

🛠️ Tech Stack

Backend

Technology	Purpose	Version
	Core Language	3.11+
	Web Framework	0.104.1
	ML Model	2.0.3
	ML Tools	1.3.0+
	Data Processing	2.0.0+
	Numerical Computing	1.24.0+

Frontend

Technology	Purpose	Version
	UI Library	18.3.1
	Build Tool	6.0.11
	Styling	3.4.17
	Data Visualization	2.15.0
	HTTP Client	1.7.9

---

📁 Project Structure

Cancer-Risk-Simulator/
├── 📂 backend/                      # Python FastAPI backend
│   ├── main.py                      # FastAPI application & endpoints
│   ├── train_model.py               # XGBoost model training script
│   ├── model.pkl                    # Serialized trained model
│   ├── requirements.txt             # Python dependencies
│   ├── The_Cancer_data_1500_V2.csv  # Training dataset (1,500 records)
│   └── __pycache__/                 # Python cache
│
├── 📂 frontend/                     # React + Vite frontend
│   ├── 📂 src/
│   │   ├── 📂 components/
│   │   │   └── RiskCalculator.jsx   # Main component (408 lines)
│   │   ├── 📂 assets/               # Static assets
│   │   ├── App.jsx                  # Root component
│   │   ├── main.jsx                 # Entry point
│   │   ├── App.css                  # Component styles
│   │   └── index.css                # Global styles + Tailwind
│   ├── 📂 public/                   # Public assets
│   ├── index.html                   # HTML template
│   ├── package.json                 # Node dependencies
│   ├── vite.config.js               # Vite configuration
│   ├── tailwind.config.js           # Tailwind configuration
│   ├── postcss.config.js            # PostCSS configuration
│   └── eslint.config.js             # ESLint configuration
│
└── README.md                        # This file

---

🚀 Installation

Prerequisites

• Python 3.11+ (Download)
• Node.js 18+ and npm (Download)
• Git (Download)

Clone Repository

git clone https://github.com/yourusername/Cancer-Risk-Simulator.git
cd Cancer-Risk-Simulator

Backend Setup

1️⃣ Navigate to backend directory

cd backend

2️⃣ Create virtual environment

python3 -m venv .venv
source .venv/bin/activate  # macOS/Linux
# .venv\Scripts\activate   # Windows

3️⃣ Install dependencies

pip install -r requirements.txt

4️⃣ Train the model (optional - model.pkl included)

python train_model.py

Output:

✓ Model loaded successfully!
Accuracy: 0.9367
ROC-AUC Score: 0.9412

5️⃣ Start backend server

uvicorn main:app --reload --host 0.0.0.0 --port 8000

✅ Backend running at: http://localhost:8000
📖 API Docs: http://localhost:8000/docs

Frontend Setup

1️⃣ Navigate to frontend directory (new terminal)

cd frontend

2️⃣ Install dependencies

npm install

3️⃣ Start development server

npm run dev

✅ Frontend running at: http://localhost:5173

---

💻 Usage

Running the Application

1. Start Backend (Terminal 1):

   cd backend
   source .venv/bin/activate
   uvicorn main:app --reload --port 8000

2. Start Frontend (Terminal 2):

   cd frontend
   npm run dev

3. Open Browser: Navigate to http://localhost:5173

4. Adjust Health Metrics: Use sliders to input your health data

5. View Results: See real-time risk prediction and analysis

Building for Production

Frontend:

cd frontend
npm run build
npm run preview  # Preview production build

Backend:

# Use production ASGI server
uvicorn main:app --host 0.0.0.0 --port 8000 --workers 4

---

Parameter Constraints

Parameter	Type	Range	Description
age	int	18-100	Age in years
gender	int	0-1	0=Female, 1=Male
bmi	float	10-50	Body Mass Index
smoking	int	0-1	0=No, 1=Yes
genetic_risk	int	0-2	0=Low, 1=Medium, 2=High
physical_activity	float	0-25	Hours per week
alcohol_intake	float	0-10	Drinks per week
cancer_history	int	0-1	0=No, 1=Yes

---

🧠 Model Details

Training Data

• Dataset: Real cancer medical records
• Size: 1,500 patient records
• Features: 8 health metrics
• Target: Binary classification (Diagnosis: 0=No Cancer, 1=Cancer)

Model Architecture

• Algorithm: XGBoost Classifier
• Tree Method: Histogram-based (optimized for CPU)
• Hyperparameters:
  • n_estimators: 150
  • max_depth: 6
  • learning_rate: 0.1
  • eval_metric: logloss

Performance Metrics

Metric	Score
Accuracy	93.67%
ROC-AUC	0.9412
Precision	~0.92
Recall	~0.94

Feature Importance

Top factors influencing predictions:

1. 🧬 Genetic Risk - Highest importance
2. 👤 Age - Strong predictor
3. 🚬 Smoking Status - Significant impact
4. 📋 Cancer History - Critical factor
5. ⚖️ BMI - Moderate importance
6. 🏃 Physical Activity - Protective factor
7. 🍷 Alcohol Intake - Risk factor
8. ⚥ Gender - Baseline risk difference

Risk Level Categories

• Low: < 25% probability (🟢 Green)
• Moderate: 25-50% (🟡 Yellow)
• High: 50-75% (🟠 Orange)
• Very High: > 75% (🔴 Red)

---

⚠️ Disclaimer

IMPORTANT MEDICAL DISCLAIMER

This application is for educational and research purposes only. While the model achieves 93.7% accuracy on test data:

• ❌ NOT a substitute for professional medical advice
• ❌ NOT a diagnostic tool for clinical use
• ❌ NOT validated for actual medical decision-making

Always consult qualified healthcare professionals for:

• Medical diagnoses
• Treatment plans
• Health screenings
• Cancer risk assessments

The predictions are based on statistical models and should be interpreted as educational demonstrations of machine learning capabilities.

---