GraphDay Stockholm - Levaraging Graph-Technology to fight Financial Fraud
TO FIGHT FINANCIAL FRAUD
Director Field Engineering
Retail Banking First-Party Fraud!
Opening many lines of credit with no intention of !
paying them back!
Causing High Impact
• Tens of billions of dollars lost every year
• 25% of total consumer credit charge-oﬀs
in the United States.(2)
• 10% to 20% of unsecured bad debt at
U.S. and European banks is misclassiﬁed,
is actually ﬁrst-party fraud.(3)
(1) Experian: http://www.experian.com/assets/decision-analytics/white-papers/ﬁrst-partyfraud-wp.pdf!
(2) Experian: http://www.experian.com/assets/decision-analytics/white-papers/ﬁrst-partyfraud-wp.pdf!
(3) Business Insider: http://www.businessinsider.com/how-to-use-social-networks-in-the-ﬁght-against-ﬁrst-party-fraud-2011-3!
Detec%ng Fraud Rings
123 NW 1st
123 NW 1st
Person A! Person B!
Location A! Location B!
Phone Number Duplicate Use
Fraud Demo – Part I (generic)!
• Fraud scenario covering Retail Fraud use cases!
• Data set contains operational data!
• Constant data load –> injecting fraud cases -> generate alerts!
• Capability to export data of detected fraud for further investigation!
Why using GraphDB / Neo4j for Fraud Detection?!
• Graphs are intuitive to understand!
• Schema free - > Flexibility!
• Nodes can vary depending on time / usage / semantic!
• Adopt dynamic changes!
• Agile Development!
• High productivity and rapid implementation !
• No “RDBMS-waterfall-high-investment-trap” !
• Taking advantage of the full value of connected data and data relationships!
• Traversing the graph compared to self joins in RDBMS!
• Near real time response times!
• Preventing fraud rather than detecting after the fact!
• Usage scenario Fraud Analyst: !
• Potential fraud case detected!
• Enriched with data from various sources containing data on fraud suspect!
• Trigger human and/or automated reactions!
Fraud Demo – Part II
DB2, HANA …)!
(E.g BI Tools such as !
Tableau, Qlik, BO,
Using Neo as the foundation of a fraud solution in
Step 1: Set up Data Integration!
Step 2: Visualize Data in BI Tool!
• Fraud as one use case to provide full value of connected data within the
• Neo4j as the foundation to do 360 degree fraud detection and
• Neo4j to extend your existing environment while protecting your
• Neo4j provides best value integrated in the entire environment!
• Neo4j as the foundation for generating real time alerts to trigger
automated or manual interventions!