Let's Connect!
I enjoy connecting with like-minded professionals passionate about technology, strategy, and impact. Feel free to reach out!
Chicago, IL
(312) 871-8022
k.teckchandani1703@gmail.com
Real-Time Fraud Detection System
Introduction
The Real-Time Fraud Detection System is a Python-based application that simulates banking transactions, detects suspicious activities through business rules and optional machine learning, and visualizes live analytics via a Streamlit dashboard. It brings together data simulation, anomaly detection, and interactive dashboards, creating an end-to-end, realistic fraud monitoring experience.
Problem Statement
Financial institutions face losses due to fraudulent transactions that aren't caught in time. Static systems fail to adapt quickly to new fraud patterns. There's a need for a transparent, explainable, and real-time fraud detection system that integrates rules, machine learning, and analytics in one streamlined pipeline.
Objectives
- Generate realistic transaction data with geolocation info.
- Detect suspicious patterns using configurable rules and optional ML.
- Explain alerts in plain English for analyst review.
- Visualize fraud trends and geo-distribution in real-time.
Architecture
💳 Transaction Simulation (Faker + Geopy)
⬇️
📜 Rule-Based Checks
⬇️
🤖 ML Scoring (Isolation Forest - optional)
⬇️
🗄 Save to SQLite Database
⬇️
📊 Streamlit Dashboard (Live Analytics + Geo Map)
Key Features
- Fraud rules: High amount, transaction frequency, IP/location mismatch, blacklist check.
- Machine Learning: Isolation Forest anomaly detection.
- Geo-Visualization: Map transactions using latitude/longitude lookups.
- Explanations: Optional GPT-based alert explanation.
Impact
- Proactive Detection: Flags anomalies instantly.
- Explainability: Gives human-readable reasons for alerts.
- Visualization: Helps fraud analysts spot trends in real time.
Conclusion
This project showcases an integrated fraud detection pipeline, from simulation to storage, detection, and visualization, making it a perfect demo of real-world fintech fraud monitoring techniques.