Johnpaul10j/fraudguard-africa

# 🛡️ FraudGuard Africa **Real-time Mobile Money Fraud Detection System** FraudGuard Africa is an end-to-end machine learning system designed to detect fraudulent mobile money transactions in real-time. The project focuses on the African fintech ecosystem and is built to be practical, explainable, and production-ready. ## 🚀 Live Demo Try the live application here: **FraudGuard Africa Live Demo** --- ## 📌 Project Overview Mobile money fraud is a growing problem across Africa. This project uses machine learning to detect suspicious transactions with high accuracy while keeping false positives under control. **Key Features:** - Real-time fraud prediction - Feature engineering tailored for mobile money transactions - FastAPI backend for production use - Modern Streamlit dashboard for easy testing - Trained on the PaySim mobile money dataset --- ## 🛠️ Tech Stack - **Language:** Python - **Machine Learning:** XGBoost - **Backend:** FastAPI - **Frontend:** Streamlit - **Data Processing:** Pandas, NumPy, Scikit-learn - **Model Persistence:** Joblib --- ## 📁 Project Structure ```bash fraudguard-africa/ │ ├── data/ # Dataset ├── models/ # Trained model files ├── notebooks/ # Exploration & training notebooks ├── src/ │ ├── main.py # FastAPI application │ └── dashboard.py # Streamlit dashboard ├── logs/ # Prediction logs ├── requirements.txt └── README.md ``` ## How to Run the Project ### Dataset The dataset is not included in this repository because of its size. Download the PaySim dataset from Kaggle: kaggle.com Place the CSV file inside the `data/` folder. 1. Clone the repository ```bash git clone github.com cd fraudguard-africa ``` 2. Create virtual environment ```bash python -m venv venv venv\Scripts\ac …