Financial Risk Monitor

Batch financial risk scoring and anomaly detection pipeline using Python, SQL, and Isolation Forest.

Executive Summary

This project analyzes 300 synthetic transaction records and produces a scored risk queue for analyst review. It combines rule-based controls with Isolation Forest anomaly detection to identify transactions that deserve escalation.

Credibility Note

Kafka and AWS are documented only as production architecture designs in docs/architecture/. They are not implemented in this local demo.

Results

Metric	Value
Transactions processed	300
Transactions flagged for escalation	24
Isolation Forest precision	0.81
Isolation Forest recall	0.74

Risk Tier Distribution

Risk Tier	Count
Low	242
Medium	57
High	1

How To Run

pip install -r requirements.txt
python scripts/risk_monitor.py

Architecture Pipeline

Synthetic transaction data
    -> Rule-based risk indicators
    -> Velocity and region risk scoring
    -> Isolation Forest anomaly detection
    -> Escalation queue output
    -> SQL monitoring views