Students Utilize Big Data to Help Solve Real-World Problems

2018-06-26T20:58:26+00:00 June 13th, 2018|

ZM Financial Systems Partners with Duke University’s Fuqua School of Business to Overcome Challenges Inherent to the Modern Data Environment

For six weeks, a group of 10 students enrolled in Duke University Fuqua’s School of Business Master of Quantitative Management (MQM) program set out to ensure the analytical theories they are learning can help financial institutions withstand another large economic downturn.

Since the financial crisis of the last decade, there are greater requirements for financial institutions to meet recent regulatory requirements, and real-world financial modeling is key to success. Right after the crisis, the Dodd-Frank Stress Test and now, more recently, the estimation of allowance for loan and lease losses (FASB’s Current Expected Credit Loss), both deal with forecasting defaults and, therefore, capital adequacy over the future economic scenarios. For financial institutions, the challenge is in the data.

The students partnered with Butch Miner, Co-Founder, ZM Financial Systems (ZMFS), a financial software company located in Cary, N.C., as part the MQM program’s capstone project: students and a faculty advisor work with an organization to solve specific problems utilizing analytic techniques. They focused on modeling two of the largest types of loans financial institutions offer: home and business loans. The MQM students were then divided into two groups, one for each loan type.

The residential loan group were directed by Mr. Miner to perform the following analysis:

  • Download large data sets of residential loan performance data from Fannie Mae and Freddie Mac;
  • Analyze the data to find variables that could impact a borrower’s ability to pay;
  • Derive a statistical model to predict defaults under different economic scenarios; and
  • Test the model against history to see the predictive power.

The business loan group began their project in a slightly different way. Because business loans do not have a publically-accessible data set like Frannie and Freddie, the team had to obtain their information from public corporate bonds. The business loan group then had to:

  • Retest the famous Altman Z-score (regression model using financial ratios) against the recent data set; and
  • Ascertain if the updated Altman Z-score is still useful in predicting business defaults.

“Both groups were successful in applying the Finance and Big Data knowledge from their Fuqua program in a real-world setting,” said Salman Azhar, Ph.D., Faculty Member, Duke University’s Fuqua School of Business. “This partnership gave Duke students an opportunity to put their learnings into practice by finding the most suitable tools and techniques for solving the business world problems. The students used cutting-edge Big Data algorithms that were useful for ZMFS.”

“I was very impressed with the analysis and student presentations,” said Butch Miner, Co-Founder, ZMFS. “The findings were consistent and heading toward results we have obtained with more time and study. The groups have definitely learned the foundations of Big Data analysis that will serve them well in their future careers.”


About Duke University Fuqua School of Business MQM Business Analytics:
Designed for college graduates with strong quantitative backgrounds, this 10-month master’s program provides training in analytics and communication, all within a specific context—finance, marketing, forensics, or strategy. More information can be found at