Demo of using the graph data science library (Neo4j GDS 1.6) on a simulated fraud data set with Bloom. You'll need Neo4j 4.0+ and GDS 1.5+, available on our download center.
This demo is built to work with a PaySim dataset generated from @voutilad's PaySim demo. You can grab a small version of the dataset here as a Neo4j dump file, or generate your own. My demo was run on a graph with ~1.5M nodes and ~2M relationships, on an instance with 16G RAM / 4 vCPUS.