Graph Databases applied to Fraud Detection
How graph databases can help to prevent fraud, a problem that is estimated to cause 16 billion $ in annual losses to the bank sector.
Nfq Innovation Team attended an event about Graph Databases organized by Neo4j. The talks were focused on use cases related to Fraud Detection, and how graph databases can help prevent a problem that is estimated to cause 16 billion $ in annual losses to the bank sector, 4 billion $ in annual losses to the ecommerce sector and 80 billion $ in annual losses to the insurance sector only in the US.
Bill Brooks, Spain Country Manager for Neo4j, opened the event with an introduction about the main features and history of Graph Databases, pointing out that already in 18th century the famous mathematician Euler was confronted with the problem known as “Seven bridges of Königsberg”, which can only be solved using Graphs Theory.
Traditional databases don’t support multiple relationships well, so the solution is graph databases.
After Bill’s talk, Jesús Barrasa, Developer Relations of Neo4j, explained in detail how to deal with some real fraud cases in ATMs and rewards programs.
Fraud patterns are detected in a more efficient way using and modeling graph databases rather than traditional pattern discovery techniques.
With Neo4j you can create very powerful queries to look for patterns with few lines of code. Example: in a database with 1 million nodes and 4 million relations, the execution time of the query is 93ms.
Miguel Fiandor, ICIJ Web developer, told us about the role that Neo4j played in the “Panama papers” leak to detect and track companies and individuals that committed fraud.
Finally, Josep Tarruella, co-founder of Graph Everywhere, a start-up company focused on helping to develop and deploy graph database projects at scale, showed us with some pedagogical examples how to query Neo4j using Cypher. He also gave the attendees some tips to convince C-level executives for the adoption of Graph Databases, using some quick wins (what he called the “Wow! Effect”).
In Nfq we are aware of the importance of Graph Databases in the Finance and Insurance sectors and we are working to adopt Neo4j in our product & solutions portfolio. We would like to thank Neo4j team for the great meeting and expect to collaborate with them in the future.