Fraud Detection using Data Science at Ticketmaster

With the rise of chip-and-pin for in-person credit card transactions, online fraud has increased dramatically. For Ticketmaster, this fraud manifests itself either through a fraudster opening a new account with a stolen credit card, or with a fraudster taking over an existing account and using the credit card on file to purchase a ticket and then transfer the ticket out of the Ticketmaster ecosystem. The models that the data scientists have rolled out into production require extensive open-source infrastructure to provide fraudulence scores for every transaction in real time. This talk will describe that infrastructure in detail, including the extensive use of Kafka Streams, Spark, Terraform, Java, Python, and Kubernetes, as well as the packages used to build the model itself and some of the innovations used in the modeling itself.

About Our Speaker

Mark Roden is an Accomplished Data Scientist at Ticketmaster, where he is the data science lead for the fraud group. In this capacity, he leads a team of data scientists to find the needles in the haystack that are the fraudulent transactions that lead to chargebacks from fans. Mark has worked at Ticketmaster for several years, where he has lead efforts in managing traffic bursts into ticket sales, recommendations across the entire Ticketmaster and Live Nation customer base, and abuse prevention from bots and malicious actors on the site.