University of Michigan mechanical engineering students forged connections between data science, mechanical engineering, and the manufacturing industry through a new course, Data Science for Manufacturing Quality Control (ME401).
The course, offered for the first time during the fall 2024 term, was taught by Chenhui Shao, associate professor of mechanical engineering. Open to undergraduate and graduate students, ME401 exposes students to data science concepts they will likely encounter in the manufacturing industry through an immersive experience with real industry data.
“Manufacturing generates big data from the machines, systems, and factories, but people often have a hard time using the data efficiently for decision-making,” Shao said. “Part of the reason is that the manufacturing workforce – the engineers – often do not have the data science background to be able to use the data effectively.”
Data science is a broad, interdisciplinary field that combines programming, statistics, mathematics, and computer science to draw conclusions and insights from large data sets. For mechanical engineers, this knowledge can help predict, visualize, and address performance challenges they may encounter in manufacturing processes and systems.
Despite the growing prevalence of data science in manufacturing and design industries, Shao said current college-level coursework often does not integrate principles of data science within the context of mechanical engineering. The lack of data science exposure can leave a gap in students’ professional skill sets that Shao aims to bridge with ME401.
Rather than following a traditional course plan, he designed the course to simulate many elements of the industry environment. His approach employs a series of instructional choices meant to expose students to data science through firsthand experience with real-world data.
“Because this is a 400-level course, I purposely made it less theoretical and more hands-on,” Shao explained.
ME401 focuses on teaching students the fundamental principles and applications of data science through the use and implementation of tools that are commonly used in industry settings, such as Python, a high-level programming language.
“In this class, it’s mostly through Python programming,” Shao said. “It’s a coding heavy class. I was not surprised that many students had no, or minimal, programming experience at the beginning of the semester.”
Instead of being assessed through exams, students advanced their Python skills throughout the semester by completing homework problems and labs in which they were given sample programming codes to modify.
At the end of the semester, students were tasked with working in teams to create and execute a final project to demonstrate their understanding of data science concepts. General Motors sponsored the final projects for the fall 2024 term, allowing students to work with real data provided by the company.
“For these projects, the students have 100% flexibility with what they do with the data,” Shao said. “They have three project topics to choose from, and then they work on the topic they choose and can implement their own ideas into the projects.”
Students explained that Shao’s efforts to make the course an experiential learning opportunity made it an especially valuable piece of their educational experience at U-M ME.
“By far, my favorite aspect of the course is the hands-on experience coming from the signal processing and data analysis of real-world datasets collected and generated in Professor Shao’s lab,” said Iago Alves Pereira, a third-year mechanical engineering PhD student.
Pereira’s three-person team for the final project was one of two teams selected to receive a “best project” award after a mock industry review judged by Shao and Guangze Li, a Senior Researcher at General Motors.
Ashley Goodrich, a senior mechanical engineering major, said she was initially drawn toward the course because she felt that the subject matter would expand her skill set as an engineer.
“I took ME401 because I was searching for a technical elective and felt it was a unique topic that would give me new analysis and statistical tools to better understand data and how to use evidence to support engineering decisions,” she said.
Like Pereira, Goodrich’s team was selected as a “best project” award winner by Shao and Li at the conclusion of the course.
ME401 stands to help better prepare ME students for careers in a field where the presence of data science and machine learning technologies is constantly growing. After collaborating with Shao to judge students’ final projects, Li said the course demonstrates the promising future of the mechanical engineering field.
“The integration of data science in manufacturing quality control is a game-changer,” he explained. “Seeing students work with real-world data and the latest technologies gives me great confidence in the future of this field. I was thrilled to see students not only using the machine learning technologies they learn in class, but also actively seeking knowledge from outside sources.”