Some criminals have started to focus on the Value Added Tax in Europe to defraud European governments of millions of Euros.
These transactions involve more than two companies and several seemingly unrelated people. The notion is to subvert traditional fraud detection measures, which are based in statistical databases.
So what would be better at detecting a new kind of fraud? A new kind of database like the graph database employed by FlockData.
It is possible to find evidence of the collusion indicative of fraud with traditional database setups. But only if you know exactly where to look. Fraud is really a relationship-based crime, where the damning evidence will be found in the connections among people, corporations, and their transactions. This is the kind of thing for which graph databases were built.
Linkurious, a graph database visualization partner of FlockData, laid out a detailed outline describing how to use graph technology and visualizations to identify potential cases of fraud.
What differentiates the graph database from classical methods of fraud detection is the ability to analyze across various data sources and profiles.
Of course, it still takes a skilled data analyst experienced with fraud data to discern real suspicious threats from cases that are just incidentally suspicious. The Linkurious visualizations go a long way in helping the data analyst do that.
Per Linkurious, “We need humans to analyse the information before decisions are made. Graph visualization solutions like Linkurious help data analysis experts investigate suspicious cases: they can decide whether a case is not suspicious or further investigate the real cases.”
The use case explained here that uncovered VAT fraud utilized Linkurious visualizations atop the Neo4j graph database, the exact setup we at FlockData use.