Just a couple of days ago, we hosted a Fraud Detection webinar. We chose to focus on credit card fraud to illustrate how Hume can help detect fraud faster. Our Director of Product, Esther Bergmark walked you through an example of a credit card company investigator looking into suspicious, fraudulent developments. Let me share what we covered at the webinar.
Introduction to Hume
GraphAware is a proud creator of Hume - a graph-powered fraud-detection and insights engine sitting on top of Neo4j. Hume is the missing link between Neo4j and the end user, which has the capacity to bring the power of graphs to your data scientists, analysts, and data-savvy business users within days.
Among the features of Hume that are incredibly powerful when detecting fraud are:
- Highly responsive graph visualization - Hume’s visualisation capabilities are fast and intuitive thanks to fast data rendering and a short time to query.
- Temporal and Geospatial views, allow you to see how your data changed over a specific period of time and enable you to see the data on a map.
- Previews help solve the exploding network problems.
- Perspectives, which are a manifestation of role-based access control, restrict access to certain parts of your graph to specific users.
- Snapshots allow you to save the current view of your canvas and share it with your colleagues, making collaboration easy.
- Alerting allows you to set up email alerts that notify you every time a new event occurs.
Payment Card Skimming Fraud
Numerous kinds of fraud need to be detected; we focused explicitly on payment card skimming fraud to show how graphs and graph-powered insights engines like Hume can help do so. Most commonly, this type of fraud occurs after a person’s credit card is skimmed, and its details are copied as the unaware customer uses their credit card on a machine that is actually a skimming machine. In 2019, payment card fraud accounted for 28.65 billion USD losses worldwide, and according to a report by the European Union, the number seems to be on a continuous rise.
In the demo part of the webinar, we walked you through a fictitious use case where a credit card holder reported an issue with their card. We used Hume to look into how the card has been used, and with one of our pre-canned actions, we quickly saw which transactions were fraudulent. We then leveraged temporal and geospatial analysis to see when and where the fraudulent transactions happened, and even more - we were able to find the store in which the credit card was most probably skimmed. Doing so allowed the investigator in our use case to find other potential credit card fraud victims by that store, thus preventing further fraud from happening.
Watch the recording from the webinar, including the demo:
Wondering how Hume could help you detect fraud? Request a demo and speak to one of our industry experts.