For a Coding-Intensive Course, Generative AI Provides a Lift
For many undergraduate business students, Introduction to Data Mining for Business (STAT 331) is the first higher-level statistics course they’ll take. Particularly for students majoring in business analytics or MIS, this course sets the foundation for more intensive work in coding and programming.
During winter and spring terms, students enrolled in this course, taught by Anna Devlin, PhD, associate clinical professor of decision sciences and MIS, were writing a guide for future classes as they worked to grasp the course material themselves.
As a Teaching Fellow with LeBow’s Center for Innovation in Teaching and Learning, Devlin sought to integrate generative AI (Gen AI) into her coursework, and together with her students, she explored using Gen AI programs to write code — a major area of application for Gen AI as the technologies have grown in popularity.
“I had seen a big dichotomy among students: some were extensively using ChatGPT for things like note taking, study preparation or coding and others who never used it before or tried and found it frustrating,” Devlin says. “I could see this divide spill over into students’ abilities to keep pace in the course and complete assignments efficiently.”
For this course, which requires using either R or Python programming languages, Devlin sought to have students hone their skills in writing specific prompts for use with Copilot, Microsoft’s Gen AI tool.
“To effectively execute data mining, you need some kind of programming language. You could understand conceptually what needs to be done but then get stuck with syntax,” she says. “In class, I encouraged students to give very specific instructions and data set descriptions to generative AI. This helped them get from knowing what analysis steps needed to be implemented to actually having executable R or Python code.”
Those steps fed into a “playbook” for the course, devised jointly by Devlin and several of LeBow’s Generative AI Student Fellows and intended to help future groups of students with coding demands.
One of the fellows, Ibraheem Ahmad, BS management information systems ’25, had taken STAT 201 and 202 with Devlin and had prior experience using Gen AI programs.
“It was incredibly rewarding to see how all the previous concepts tied together,” he said. “I feel that if I were to work on a data science project now, I would have a solid foundation to contribute effectively without feeling overwhelmed.”
Sophia Vincenty, BS management information systems and business analytics ’25, noted that the focus on business applications for data mining and the rigor of the final project made the course her favorite so far in her studies at Drexel.
“We routinely kept a target business question in mind, which shaped why and how to approach the data mining process in an appropriate way,” she says. “The basis of the course was to formulate and interpret statistical programming models based on target and predictor variables, which allowed us to critically think of consumer and market behaviors that indirectly lead to business opportunity.”
During fall term, Devlin plans to use generative AI in an operations and supply chain course on sustainability, and she’ll also begin to pilot a coding “bootcamp” for undergraduates that will draw on the playbook created earlier this year.
“Everybody should be exposed to this for coding purposes, because coding is how you automate things,” she says. “Now that these students gotten comfortable doing it here, they’ll apply it to other things.”
