mauree155/FraudWatch-Africa

🚨 FraudWatch Africa – A machine learning-powered app using FastAPI + Streamlit to detect and analyze fraudulent financial transactions in Kenya # FraudWatch Africa: Detecting Fraud in Mobile Money Transactions with Unsupervised Learning ## Table of Contents 1. Project Background - Mobile Money in Africa - The Fraud Detection Challenge 2. Project Goal 3. Key Features 4. Methodology - Exploratory Data Analysis (EDA) - Data Preprocessing - Modeling 5. Results and Discussion 6. Dashboard & Deployment 7. Tools & Technologies 8. Conclusion 9. Future Work 10. How to Run the Project 11. Acknowledgments ## Project Background ### Mobile Money in Africa Mobile money has transformed financial inclusion in Africa. Services like **M-Pesa (Kenya)**, **MTN Mobile Money (Uganda)**, and **Airtel Money (West Africa)** allow millions of people to send money, pay bills, and manage their finances without relying on traditional banks. With over **300 million active users in Sub-Saharan Africa**, mobile money platforms are now the backbone of everyday transactions. However, this rapid growth also introduces **security challenges**: - Limited regulatory oversight - High transaction volumes - The anonymity of mobile wallets Together, these factors make mobile money ecosystems a **prime target for fraudsters**. Common fraud tactics include: - SIM swaps - Account takeovers - Fraudulent transfers ### The Fraud Detection Challenge Fraudulent transactions are notoriously **difficult to detect** because they rarely follow predictable patterns. Traditional supervised machine learning approaches require **labeled fraudulent data**, which is often scarce or unavailable. To address this challenge, this project leverages **unsupervised learning**, where the model learns to identify **outliers** that deviate from normal transaction behavior — a promising approach in fraud detection for data-scarce environments. ## Project Goal This project aims to design a **scalable, real-time fraud detection system** tailored to mobile money platforms in Africa. Ke …