PometryDocs
GitHub

Intelligence Analyst Tutorial

From raw data to actionable intelligence in 30 minutes.

Learn how to use Raphtory to detect fraud rings, trace money flows, uncover Ultimate Beneficial Owners (UBOs), and generate investigation-ready narratives - without writing complex code.

What You'll Accomplish

  1. Ingest transaction data from your existing data warehouse
  2. Detect coordinated account creation (synthetic identity fraud)
  3. Trace multi-hop money flows across temporal windows
  4. Score risk dynamically based on behavioral patterns
  5. Export investigation packages for case management

Time: 30 minutes
Prerequisites: Access to transaction/entity data. Python basics helpful but not required.

1. Load Your Data

Connect to your transaction data. Raphtory works with any tabular source - CSV, Parquet, SQL, or direct warehouse connectors.

2. Detect Coordinated Account Creation

Fraudsters often create multiple accounts simultaneously. Find accounts born within the same hour that immediately transact with each other.

Real-World Calibration: The 5-account threshold and 1-hour window are starting points. Tune based on your false positive rate.

3. Trace Multi-Hop Money Flows

Follow the money across multiple hops while respecting time ordering. This is where temporal graphs shine - you can't move money backward in time.

4. Dynamic Risk Scoring

Score each account based on temporal behavioral patterns - not just static attributes.

5. Export Investigation Package

Generate outputs ready for your case management system.

Advanced: Unmasking Ultimate Beneficial Owners (UBOs)

Trace ownership and control relationships across corporate structures:

Next Steps