This is a random data generation program for the "Financial Crime Discovery using Amazon EKS and Graph Databases" article by Zahi Ben Shabat, in which he describes how to use RDFox and its reasoning capabilities to detect suspicious patterns in banking data.

The program is meant to simulate a record of bank transactions between individuals. It first generates a configurable number of parties, some of which are marked as "suspicious", and then a configurable number of transactions between them. Each transaction's "originator party" and "beneficiary party" are selected entirely randomly from the pool, and the transaction amount is also random. The program outputs RDF files in Turtle syntax.

RDFox can find transactions involving "suspicious" individuals and follow "transaction chains" originating from them to find instances where money is being funneled from one criminal to another through a number of seemingly legitimate accounts (also known as money laundering).

RDFox is the world's most performant knowledge graph and reasoning engine. To try it for yourself, request a free trial here.

For Help, run:

python main.py --help

Usage:

python main.py -o data -tc $NUM_OF_MSG -pc $NUM_OF_PARTIES -mdb $MAX_DAYS_BEFORE -fc $NUM_OF_FILES -thc $THREAD_COUNT

This will create or overwrite the files in the 'data' folder.

Generate 10,000 transactions in 10 files, 1000 transaction per file

python main.py -o data -tc 100000 -pc 1000 -mdb 3 -fc 100 -thc 10

Generate 5,000,000 transactions in 100 files, 50,000 transaction per file

python main.py -o data -tc 5000000 -fc 100