View on GitHub

Fraud Detection Platform

An end-to-end machine learning platform for detecting fraudulent transactions in real-time.

🛡️ Fraud Detection Platform

Real-time Fraud Detection System powered by Machine Learning

A production-ready, end-to-end fraud detection platform built with FastAPI, Docker, and Machine Learning. Detect fraudulent transactions in real-time with monitoring, observability, and a beautiful dashboard.

📋 Table of Contents

✨ Features
🏗️ Architecture
🚀 Quick Start
📊 Dashboard
📁 Project Structure
⚙️ Configuration
🔧 Development
📈 Monitoring
🤝 Contributing
📄 License

✨ Features

Core Features

🔍 Real-time Fraud Detection - Detect fraudulent transactions in milliseconds
🤖 ML Models - XGBoost, LightGBM, CatBoost
📊 Data Pipeline - ETL pipeline with data generation and validation
🗄️ Feature Store - Lightweight feature management

Infrastructure

🐳 Dockerized deployment
📈 Prometheus & Grafana Monitoring
📊 Streamlit Dashboard
🧪 MLflow Experiment Tracking

API & Integration

🚀 FastAPI
📝 Swagger/OpenAPI
🔄 Kafka Streaming
🗄️ PostgreSQL

🏗️ Architecture

┌─────────────────────────────────────────────────────────────────────┐
│                     Client Applications                             │
├─────────────────────────────────────────────────────────────────────┤
│                  Streamlit Dashboard (8501)                         │
├─────────────────────────────────────────────────────────────────────┤
│                         FastAPI (8000)                              │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐                  │
│ │   /detect    │ │   /history   │ │    /stats    │                  │
│ └──────────────┘ └──────────────┘ └──────────────┘                  │
│ ┌───────────────────────────────────────────────────────────────┐   │
│ │             FraudService (Business Logic)                     │   │
│ └───────────────────────────────────────────────────────────────┘   │
├─────────────────────────────────────────────────────────────────────┤
│                  Machine Learning Engine                           │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐                  │
│ │ LightGBM     │ │   XGBoost    │ │  CatBoost    │                  │
│ └──────────────┘ └──────────────┘ └──────────────┘                  │
├─────────────────────────────────────────────────────────────────────┤
│                      Data & Storage                                │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐                  │
│ │ PostgreSQL   │ │    Kafka     │ │    Models    │                  │
│ └──────────────┘ └──────────────┘ └──────────────┘                  │
├─────────────────────────────────────────────────────────────────────┤
│                     Monitoring Stack                                │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │  Prometheus  │ │   Grafana    │ │  Evidently   │ |    MlFlow    | │
│ └──────────────┘ └──────────────┘ └──────────────┘ └──────────────┘ │
└─────────────────────────────────────────────────────────────────────┘

🚀 Quick Start

Prerequisites

Docker 24+
Docker Compose 2.20+
Python 3.11+

Clone Repository

git clone https://github.com/yourusername/fraud-detection-platform.git
cd fraud-detection-platform

Run Docker

docker-compose up -d

docker-compose ps

docker-compose logs -f api

Available Services

Service	URL
API	http://localhost:8000
Swagger	http://localhost:8000/docs
Streamlit	http://localhost:8501
Grafana	http://localhost:3000
PGAdmin	http://localhost:5050
Kafka UI	http://localhost:8080
MlFlow	http://localhost:5000

Test API

curl -X POST http://localhost:8000/api/v1/fraud/detect \
-H "Content-Type: application/json" \
-d '{
  "transaction_id":"TX123456",
  "user_id":"user_3851",
  "amount":1500.0,
  "timestamp":"2024-07-05T10:30:00",
  "merchant":"Amazon"
}'

curl http://localhost:8000/api/v1/fraud/stats?days=7

curl "http://localhost:8000/api/v1/fraud/history/all?days=7&limit=50"

📊 Dashboard

Streamlit

📈 Overview
🔍 Real-time Detection
📊 Analytics
👤 User Risk Profile
📉 Model Performance

Grafana

API Metrics
Model Performance
Business Metrics

📁 Project Structure

fraud-detection-platform/
├── src/
│   ├── api/
│   │   ├── routers/
│   │   ├── schemas/
│   │   └── services/
│   ├── core/
│   ├── data_engineering/
│   └── ml_engineering/
├── dashboard/
├── tests/
├── models/
├── data/
├── grafana/
├── monitoring_reports/
├── docker-compose.yml
├── Dockerfile
├── requirements.txt
└── README.md

⚙️ Configuration

Environment Variables

DB_HOST=postgres
DB_PORT=5432
DB_NAME=fraud_db
DB_USER=fraud_user
DB_PASSWORD=fraud_pass

API_HOST=0.0.0.0
API_PORT=8000
API_DEBUG=True

KAFKA_BOOTSTRAP_SERVERS=kafka:9092

MLFLOW_TRACKING_URI=http://mlflow:5000

🤖 Supported Models

Model	Status
XGBoost	✅
LightGBM	✅
CatBoost	✅

🔧 Development

Local Setup

python3.11 -m venv venv

source venv/bin/activate

pip install -r requirements.txt

uvicorn src.api.main:app --reload

streamlit run dashboard/app.py

Train Model

docker exec -it fraud_api python scripts/train_model.py

or

python scripts/train_model.py

Run Tests

pytest tests/ -v --cov=src

Generate Data

docker exec fraud_api python src/data_engineering/generator.py

docker exec fraud_api python src/data_engineering/etl_pipeline.py

📈 Monitoring

Prometheus Metrics

http_requests_total
http_request_duration_seconds
fraud_predictions_total
model_accuracy

Grafana Dashboards

API Performance
Fraud Analytics
Model Performance
System Health

🤝 Contributing

Fork repository
Create feature branch

git checkout -b feature/amazing-feature

Commit

git commit -m "Add amazing feature"

Push

git push origin feature/amazing-feature

Open Pull Request

📄 License

Licensed under the MIT License.

🙏 Acknowledgments

FastAPI
Scikit-learn
Docker
Grafana
Streamlit

👨‍💻 Author

Sobhan Alizadeh

AI Engineer · Builder of Intelligent Systems · RAG & LLM Enthusiast

⭐ Support

If you like this project, please give it a ⭐ on GitHub.

Built with ❤️ for Fraud Detection.