Salar Fattahi, U-M Industrial and Operations Engineering (IOE) assistant professor, has spearheaded the creation of a new data analytics course that combines optimization techniques and data science with real-world case studies.
“Generally, in most data analytics courses, the focus is mostly on the statistical side, with optimization coming into play as a black box, a solver that can be applied without knowing how or why it works,” said Fattahi. “In contrast, our goal is to not only give the tools, but also cover why and how these optimization techniques work in practice.”
“Every week I’ll introduce a new case study, often drawing from my own research,” said Fattahi. “I’ll explain what the problem is, why do we even care about this problem and, how we can solve it.”
“For example, in many business cases data leads the process of decision making, and sometimes the scale of the data is very large, or it is corrupted with noise. This leads to several important questions: How do you select a model that captures the data generation process? How do you deal with large-scale data? What is the best kind of algorithm you can use to solve this problem efficiently?” he said.
“These are the things that we’re planning to cover in this course. So that our engineers can make more informed choices about which models and algorithm they should use to solve any particular problem efficiently.”
The course is designed and assessed around projects rather than midterms or final exams. Students are free to design their own projects and pick their own topics. It will be open to U-M IOE undergraduate and master’s students and is the second of a two-course pair that focuses on modern tools in data analytics. The first course in the pair is IOE 373, “Data Analytics Tools and Techniques.”
The new course is best suited for students who have experience in the basics of linear algebra, probability, and data science.
Salar Fattahi joined U-M IOE as an assistant professor in 2020. He completed his master’s and doctoral degrees in industrial engineering and operations research from the University of California, Berkeley. He has also received a master’s degree in electrical engineering from Columbia University. Fattahi’s research focuses primarily on the intersection between optimization, machine learning and control theory.