In this session, Steve discusses the challenges of collecting and cleansing data for fraud analysis, specifically in relation to increasing fraud losses through mobile and ATM deposits. Suspected to be related to a scam called "card popping" or "card cracking," Steve spends months collecting and cleansing the data using advanced analytics to identify fraudulent mobile devices and correlate RDI deposits with mobile and ATM channels. The results showed a correlation between young accounts and mobile deposits coming back as RDIs, leading to new policies restricting mobile deposit use for young accounts. He also used K-means cluster analysis and logistic regression to identify fraud trends and improve fraud detection. The initial goal was to save between 1.5 to $1.8 million, and the latest results from January to the end of February show a savings of $849,000 year-to-date, but there was an unexpected increase in fraud related to skimmers in the Tampa area, resulting in an additional $110,000 in savings, bringing the total savings to just under $950,000.