Data Mining Class Gives Students Real-World Experience
University of Scranton Associate Professor Yibai Li, Ph.D., is always looking for unique ways for students in his Data Mining course to apply their knowledge in an actual professional context.
In the spring of 2018, he found one such opportunity – working with Lititz, Pennsylvania-based R.W. Sauder, one of the country’s biggest egg wholesalers.
Students in the course spent the entire semester attempting to find out whether data mining could predict egg prices six weeks ahead of time. That would allow R.W. Sauder to better manage its inventory, reduce its costs and increase its profits.
By the end of the project, the class had developed a model that could predict egg prices six weeks out -- at an accuracy rate of 92.8 percent when the confidence level was above 0.57, according to Dr. Li.
“I was pleased with the students' performance, considering data mining is a difficult topic,” said Dr. Li, a member of the Operations and Information Management Department. “Typically, it requires many years of learning, sometimes a Ph.D., before someone can complete a project like this. We did it within one semester. The students learned data mining from scratch and were able to complete a real-world project by the end of the semester.”
Various Algorithms Used
To start the project, Dr. Li divided the class into five groups, with each working independently, yet comparing their results to improve each other's models. Besides gathering Sauder’s 10-year sales numbers and egg prices, the students collected some other useful information, including Avian flu incidents, severe weather incidents and temperature over the same decade.
“I think the project was done very rigorously,” Dr. Li said. “The students followed the data mining methodology they learned in the course and tried many different cutting edge algorithms and optimized the results accurately.”
The students presented their results to R.W. Sauder's President and CEO, Mark Sauder, and other company representatives. The executives found the results useful and commended the students for their efforts.
Mutually Beneficial Results
Besides providing a significant service to R.W. Sauder, the students made themselves more marketable in the process.
“This project gave me great exposure to not only polishing my data mining skills, but also in working on my presentation skills, which continues to help me immensely in my current role,” said Hurr Hamdani, a securitized products analyst at JPMorgan Chase & Co. He received his MBA from the University in 2018.
Meanwhile, Hamdani's classmates have landed positions at other top-flight companies, including Goldman Sachs.
“The students,” Dr. Li said, “got real-world project experience that they could put on their resumes.”