ATM Fraud Detection with Oracle Property Graphs

With a stream of ATM transactions in Apache Kafka, how to detect if a card has been cloned? Turn the stream into a property graph and use PGQL to detect possible frauds. Two challenges in one talk: streams to property graphs and fraud detection. Proving the power and flexibility of graph databases, the mandatory product in your Analytics toolbox!

Transcript

1. View Slide

2. View Slide

3. Copyright © 2017, Oracle and/or its affiliates. All rights reserved. |
   bit.ly/OracleACEProgram
   450+ Technical Experts
   Helping Peers Globally
   Nominate yourself or someone you know: acenomination.oracle.com
   

   View Slide

4. Vertex
   edge
   A Property Graph
   (also called node)
   

   View Slide

5. edge
   edge label
   edge properties
   edge ID
   directed edge
   vertex (node)
   vertex
   properties
   vertex ID
   a vertex can
   have a label
   

   View Slide

6. View Slide

7. Every row of a table has a fixed, identical structure
   Connections at a table level (not row)
   Nodes and edges can have any number of properties
   Connections at a node level (can be seen as row level)
   

   View Slide

8. Spain
   Italy
   John Doe
   Company A
   Company B
   Company C
   Company D
   Located in
   Located in
   Located in
   Located in
   Buys from
   Buys from
   Buys from
   Buys from
   Money laundering
   and VAT frauds
   Owns
   

   View Slide

9. Good guy
   Bad guy
   Card cloned while
   using an ATM
   Somewhere else
   

   View Slide

10. bank
    customer
    account account
    customer
    ATM
    account
    CC
    transfer money to
    owns
    owns
    owns
    get money from
    

    View Slide

11. •
    •
    •
    •
    •
    •
    The presentation was a live demo,
    screenshots aren’t as explicit as seeing it…
    

    View Slide

12. View Slide

13. View Slide

14. {
    "account_id": "a839",
    "timestamp": "2019-03-10 13:49:32 +0000",
    "atm": "24/7",
    "amount": 20,
    "location": {
    "lat": "53.7835061",
    "lon": "-1.3794149"
    },
    "transaction_id": "555aa4f6-433b-11e9-b9b9-0242ac140005"
    }
    Location of the ATM
    

    View Slide

15. Kafka
    stream
    Property Graph
    

    View Slide

16. Kafka
    stream
    Property Graph
    A Java app consume the Kafka stream and update the graph with
    the new messages received.
    

    View Slide

17. View Slide

18. View Slide

19. View Slide

20. The number of edges increase over time, it’s the Kafka
    messages being loaded in the graph
    

    View Slide

21. PGQL query, looks like SQL but has the extra GRAPH syntax
    

    View Slide

22. PGQL query, looks like SQL but has the extra GRAPH syntax
    

    View Slide

23. View Slide

24. View Slide

25. View Slide

26. View Slide

27. View Slide

28. The UK set of ATMs used is a bit too sparse over a big
    area. Too many of the randomly generated transactions
    seem to be fraud.
    

    View Slide

29. The UK set of ATMs used is a bit too sparse over a big
    area. Too many of the randomly generated transactions
    seem to be fraud.
    

    View Slide

30. View Slide

31. View Slide

32. View Slide

33. View Slide

34. View Slide

35. •
    •
    •
    •
    •
    •
    •
    more about graphs:
    Data Lineage Made Easy with Graph Database
    14:50-15:20 | Sentosa 1 (the other room)
    UKOUG Techfest 2019, Dec 2019 in Brighton
    

    View Slide