https://github.com/mrkunalsharma/fraud-detection-mlops
🚀 Production-ready fraud detection system with 99.95% accuracy, real-time API, and comprehensive monitoring
https://github.com/mrkunalsharma/fraud-detection-mlops
docker fastapi fraud-detection grafana machine-learning mlops prometheus python
Last synced: about 2 months ago
JSON representation
🚀 Production-ready fraud detection system with 99.95% accuracy, real-time API, and comprehensive monitoring
- Host: GitHub
- URL: https://github.com/mrkunalsharma/fraud-detection-mlops
- Owner: MrKunalSharma
- Created: 2025-09-20T17:00:14.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2025-09-20T19:10:05.000Z (9 months ago)
- Last Synced: 2025-09-20T19:10:29.219Z (9 months ago)
- Topics: docker, fastapi, fraud-detection, grafana, machine-learning, mlops, prometheus, python
- Language: Python
- Homepage:
- Size: 52.7 KB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
- Codeowners: .github/CODEOWNERS
Awesome Lists containing this project
README
# 🛡️ Real-time Fraud Detection System with MLOps Pipeline
[](https://www.python.org/downloads/)
[](https://fastapi.tiangolo.com/)
[](https://www.docker.com/)
[](https://prometheus.io/)
[](https://fraud-detection-mlops-msnfh7adt5ykkr6yzbmhyv.streamlit.app/)
[](https://fraud-detection-api-v5cc.onrender.com/docs)
A production-ready machine learning system for real-time credit card fraud detection, featuring automated model training, API deployment, and comprehensive monitoring.
> **🎯 Perfect for**: Data Scientists, ML Engineers, DevOps Engineers, and anyone interested in MLOps best practices
## 🌟 Key Features
- **Real-time Fraud Detection**: REST API endpoint for instant fraud prediction
- **MLOps Pipeline**: Automated data preprocessing, model training, and evaluation
- **Model Comparison**: Evaluates Random Forest, Logistic Regression, and Decision Tree
- **Production Monitoring**: Prometheus metrics and Grafana dashboards
- **Containerized Deployment**: Docker Compose for easy deployment
- **High Performance**: Sub-100ms prediction latency
## 🚀 Live Demo
### Try it Now!
🎯 **[Fraud Detection System Demo](https://fraud-detection-mlops-msnfh7adt5ykkr6yzbmhyv.streamlit.app/)**
### API Endpoints
📚 **[Interactive API Docs](https://fraud-detection-api-v5cc.onrender.com/docs)**
🔍 **[Health Check](https://fraud-detection-api-v5cc.onrender.com/health)**
## 🚀 Quick Start
### Prerequisites
- Python 3.10+
- Docker Desktop
- 8GB RAM minimum
### Installation
**Option 1: Try the Live Demo**
🚀 Visit [https://fraud-detection-mlops-msnfh7adt5ykkr6yzbmhyv.streamlit.app/](https://fraud-detection-mlops-msnfh7adt5ykkr6yzbmhyv.streamlit.app/)
**Option 2: Run Locally**
1. **Clone the repository**
```bash
git clone https://github.com/MrKunalSharma/fraud-detection-mlops.git
cd fraud-detection-mlops
```
2. **Run with Docker Compose**
```bash
docker-compose up --build
```
3. **Access the services**
- 📚 **API Documentation**: [http://localhost:8000/docs](http://localhost:8000/docs)
- 📊 **Prometheus Metrics**: [http://localhost:9090](http://localhost:9090)
- 📈 **Grafana Dashboard**: [http://localhost:3000](http://localhost:3000) (admin/admin)
## 💻 Local Development
1. **Create virtual environment**
```bash
python -m venv venv
venv\Scripts\activate # Windows
source venv/bin/activate # Linux/Mac
```
2. **Install dependencies**
```bash
pip install -r requirements.txt
```
3. **Download dataset**
```bash
python download_data.py
```
4. **Train models**
```bash
python -m src.data_preprocessing
python -m src.model_training
```
5. **Run API locally**
```bash
python run_api.py
```
## 📊 Model Performance
| Model | Accuracy | Precision | Recall | F1-Score | ROC-AUC |
|-------|----------|-----------|--------|----------|---------|
| **Random Forest** | 99.95% | 95.12% | 82.93% | 88.61% | 0.9146 |
| **Logistic Regression** | 99.91% | 91.67% | 73.33% | 81.48% | 0.8666 |
| **Decision Tree** | 99.89% | 86.21% | 83.33% | 84.75% | 0.9165 |
> **🏆 Winner**: Random Forest achieves the best overall performance with 99.95% accuracy and 88.61% F1-Score
## 🔄 API Usage
### Health Check
```bash
curl http://localhost:8000/health
```
### Fraud Prediction
```bash
curl -X POST "http://localhost:8000/predict" \
-H "Content-Type: application/json" \
-d '{
"Time": 0,
"V1": -1.359807,
"V2": -0.072781,
"V3": 2.536347,
"V4": 1.378155,
"V5": -0.338321,
"V6": 0.462388,
"V7": 0.239599,
"V8": 0.098698,
"V9": 0.363787,
"V10": 0.090794,
"V11": -0.551600,
"V12": -0.617801,
"V13": -0.991390,
"V14": -0.311169,
"V15": 1.468177,
"V16": -0.470401,
"V17": 0.207971,
"V18": 0.025791,
"V19": 0.403993,
"V20": 0.251412,
"V21": -0.018307,
"V22": 0.277838,
"V23": -0.110474,
"V24": 0.066928,
"V25": 0.128539,
"V26": -0.189115,
"V27": 0.133558,
"V28": -0.021053,
"Amount": 149.62
}'
```
### Response
```json
{
"prediction": 0,
"probability": 0.02,
"risk_level": "Low",
"message": "Transaction seems legitimate"
}
```
## 📈 Monitoring & Metrics
### 📊 Tracked Metrics
- **Prediction Analytics**: Total predictions count by type (fraud/legitimate)
- **Performance**: Prediction latency histogram and response times
- **Reliability**: API request success/error rates
- **Model Health**: Real-time model performance metrics
### 📈 Grafana Dashboards
Access comprehensive visualizations:
- **Real-time fraud detection rate**
- **Transaction risk distribution**
- **API response times**
- **System health metrics**
- **A/B testing results**
## 🗂️ Project Structure
```
fraud-detection-mlops/
├── src/
│ ├── config.py # Configuration settings
│ ├── data_preprocessing.py # Data pipeline
│ ├── model_training.py # Model training logic
│ ├── model_serving.py # FastAPI application
│ └── monitoring.py # Metrics tracking
├── models/ # Trained models
├── data/ # Dataset storage
├── docker/ # Docker configurations
├── monitoring/ # Prometheus & Grafana configs
├── tests/ # Unit tests
├── requirements.txt # Python dependencies
├── Dockerfile # Container definition
└── docker-compose.yml # Multi-container setup
```
## 🔧 Configuration
Key configurations in `src/config.py`:
- **Model hyperparameters**: Learning rates, tree depth, regularization
- **API settings**: Port, host, rate limiting, authentication
- **Monitoring intervals**: Metrics collection frequency, alert thresholds
- **Data paths**: Input/output directories, model storage locations
## 🚀 Advanced Features
### A/B Testing & Model Versioning
- Multiple model versions deployed simultaneously
- Traffic splitting for gradual rollouts
- Performance comparison metrics
- Automatic winner selection
### API Authentication
```bash
# Include API key in requests
curl -X POST "https://api-url/predict" \
-H "X-API-Key: your-api-key" \
-H "Content-Type: application/json" \
-d '{...}'
```
### Data Drift Detection
- Real-time monitoring of feature distributions
- Automatic alerts when drift detected
- Z-score based drift calculation
- Integrated with monitoring dashboard
### Load Testing Results
- Handles 120+ requests/second
- 99.8% success rate under load
- <100ms average response time
- Detailed performance documentation in `/docs/PERFORMANCE.md`
### Monitoring Endpoints
- `/metrics` - Prometheus metrics
- `/health` - Health check
- `/model-performance` - A/B test results
## 📝 Key Learnings
This project demonstrates:
- **End-to-end ML pipeline development** from data preprocessing to deployment
- **REST API design** with FastAPI and proper error handling
- **Containerization** with Docker and Docker Compose
- **Production monitoring** with Prometheus/Grafana integration
- **MLOps best practices** including CI/CD, testing, and monitoring
- **A/B testing** for model comparison and gradual rollouts
- **Data drift detection** for model reliability
## 🤝 Contributing
We welcome contributions! Here's how to get started:
1. **Fork the repository** and clone your fork
2. **Create a feature branch** (`git checkout -b feature/AmazingFeature`)
3. **Make your changes** and add tests if applicable
4. **Commit your changes** (`git commit -m 'Add AmazingFeature'`)
5. **Push to your branch** (`git push origin feature/AmazingFeature`)
6. **Open a Pull Request** with a clear description
### 🎯 Areas for Contribution
- 🐛 Bug fixes and improvements
- 📊 Additional model algorithms
- 📈 Enhanced monitoring features
- 🧪 More comprehensive tests
- 📚 Documentation improvements
## 📜 License
This project is licensed under the MIT License - see the LICENSE file for details.
## 🙏 Acknowledgments
- **Dataset**: [Kaggle Credit Card Fraud Detection](https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud) - High-quality anonymized credit card transactions
- **Inspiration**: Real-world fraud detection systems and MLOps best practices
- **Technologies**: Built with modern MLOps tools and practices
- **Community**: Thanks to the open-source community for the amazing tools and libraries
## 📧 Contact
**Kunal Sharma**
- LinkedIn: [https://www.linkedin.com/in/kunal-sharma-1a8457257/](https://www.linkedin.com/in/kunal-sharma-1a8457257/)
- Email: kunalsharma13579kunals@gmail.com
- GitHub: [https://github.com/MrKunalSharma](https://github.com/MrKunalSharma)
- Project Link: [https://github.com/MrKunalSharma/fraud-detection-mlops](https://github.com/MrKunalSharma/fraud-detection-mlops)
---
⭐ **If you found this project helpful, please consider giving it a star!**
[](https://github.com/MrKunalSharma/fraud-detection-mlops)
[](https://github.com/MrKunalSharma/fraud-detection-mlops)
[](https://github.com/MrKunalSharma/fraud-detection-mlops)