Dannywhilz001/Fintech-Fraud-Detection-Nigeria

50k Nigerian transactions | Real-time fraud rules | RFM | Cohorts | Velocity detection | Full SQL + Python # Fintech Fraud Detection & Analytics (50,000 Nigerian Transactions) Dashboard Preview Real-world Nigerian fintech fraud detection & analytics system built from scratch using a fully synthetic, realistic 50,000-transaction dataset Detects fraud using *velocity, channel risk, amount thresholds, merchant scoring* — exactly how Opay, Moniepoint & Kuda do it. ## Key Features - 50,000 realistic Nigerian transactions (2024–2025) - 5,000 customers | 200 merchants - 8+ real fraud detection rules (velocity, USSD abuse, reversal patterns) - RFM customer segmentation - Monthly active users & churn tracking - High-risk merchant exposure - All in *pure SQL + Python (pandas + SQLite)* ## Fraud Detection Rules Implemented | Rule | Detection Rate | |-----------------------|----------------| | High amount (>₦1.5M) | Very High | | USSD >₦800k | Very High | | 8+ txns in 10 mins | Extremely High | | High-risk merchants | High | ## Tech Stack - Python · Pandas · SQLite · Matplotlib/Seaborn - 100% reproducible · No API keys ## How to Run ```bash # Just open the notebook and run all cells jupyter notebook Fintech_Fraud_Analytics.ipynb Built to show recruiters: “I don’t just know SQL — I catch fraud like a Nigerian fintech pro.” Nigeria | Fintech | Fraud Detection | SQL | Data Analytics | Portfolio Project ⭐ Star if you’re building the future of African fintech.