Mastertactician23/transaction-anomaly-detector

Python-based financial transaction anomaly detection engine — 11 fraud rules mapped to East African fintech attack patterns including velocity attacks, SIM-swap, structuring, layering, and mule accounts # Transaction Anomaly Detection Engine ### A Python-based financial transaction anomaly detection system implementing 11 fraud detection rules mapped to documented East African fintech attack patterns — covering velocity attacks, structuring, SIM-swap fraud, layering, mule accounts, cross-border corridors, and KYC bypass — producing risk-scored alerts and an HTML investigation report **Author:** Kofi Asibey-Kitiabi **GitHub:** Mastertactician23 **LinkedIn:** asibey-kitiabi **Date:** July 2026 **Status:** Completed **Difficulty:** Intermediate–Advanced --- ## Threat Intelligence Context > *"Kenya recorded 4.56 billion cyber threat events in Q4 2025 — a 441% surge in three months. Mobile banking fraud cases surged 87%, driven by SIM-swap schemes, credential theft, and social engineering."* > — Communications Authority Kenya / Techweez, March 2026 > *"Eastern European syndicates including Carbanak and FIN7 have expanded operations into Kenya, Uganda, and Ghana, executing sophisticated malware attacks and unauthorised transfers."* > — International Finance Magazine / INTERPOL Africa, 2025 > *"Nigeria's Central Bank now requires all licensed financial institutions to implement real-time fraud detection systems."* > — CBN Cybersecurity Framework 2025 East Africa's 459 million mobile money accounts represent the largest mobile financial services ecosystem in the world — and one of the fastest-growing attack surfaces. This engine was built to detect the fraud patterns that are actively targeting that ecosystem right now. --- ## Table of Contents 1. Project Overview 2. Detection Rules 3. Architecture 4. Tools & Technologies 5. How to Run It 6. Synthetic Dataset 7. Sample Results 8. HTML Report 9. MITRE ATT&CK Mapping 10. Regulatory Alignment 11. Connection to Portfolio 12. Skills Demonstrated 13. Design Decisions 14. What I Would Do Differently 15. Next Steps --- ## 1. Project Overview The Transaction Anomaly Detection Engine analyses financial transaction logs …