Building a Graph Native ML Pipeline for detecting Money Mules

The Dataset

The dataset uses a synthetic p2p money network called PaySim.

While you can use the PaySim Demo Loader to bootstrap a new database, you can take a shortcut and download the provided dump file. Follow the Neo4j docs on loading dump files or pushing them to Aura.

The Bloom Perspective

To recreate the Bloom experience shown in the demo, download the provided perspective file which you can import into Neo4j Bloom.

The IPython Notebook

The heart of the demo lies in the provided notebook. In order to run it, you'll need the following pre-requisites:

1. An IPython environment, like Jupyter, with a Python3 kernel.
2. Ability to pip install packages.
3. Connection details to a Neo4j AuraDS or Neo4j Enterprise server (with the GDS v2.1.x plugin installed).
   • Make sure to update the notebook with your connection details!

The notebook may work with Neo4j Community Edition, but it's untested.