PSECU Advanced Analytics Team Honored with CUNA Excellence in Technology Award

PSECU was recently honored with an Excellence in Technology Award from the Credit Union National Association (CUNA) Councils for their project entry, “Predicting Loan Application Fraud”, in the Miscellaneous category.

The award identifies, recognizes, and honors credit unions that exemplify excellence in information technology. It showcases innovative approaches to serving members, acknowledging their universal application to credit unions nationwide, and sharing the solutions with peers.

Scott Maronic, PSECU Director of Data Enterprise Analytics & Governance, and the rest of the PSECU Advanced Analytics team, developed the project in response to a request to help improve fraud mitigation for the loan application process. The team worked closely with their colleagues in Revenue, Growth, and Lending to identify characteristics thought to be indicative of fraud. Utilizing data from PSECU’s core system, loan application platform, and fraud investigation tool they began analyzing the statistically significant characteristics of a fraudulent applicant.

“Then, we began building our predictive model to help score applicants and allow our financial crimes team to prioritize their investigations,” explained Maronic. “After training and testing, we determined our model could predict whether an applicant was fraudulent or not with 97% accuracy!”

Only two months later, they were able to produce the first targeted list to PSECU’s financial crimes team. Ultimately, it helped curb hundreds of thousands of dollars of at-risk credit lines within a week’s time. Over the next year, the expectation is that the tool will be able to deter millions of dollars’ worth of fraudulent credit lines.

The opportunity to innovate solutions in order to make PSECU more efficient is one of the best parts of working in business intelligence,” Maronic continued. “To be recognized with an award for that work is just icing on the cake. This project would not have been possible without two members of the team, in particular: John Whitehead, our Data Scientist, and Noah Hake, a Data Analyst. Their expertise made this project successful.”