In this section, Prof. Georgia Perakis describes her process for developing case studies for 15.060 Data, Models, and Decisions.
To develop a case, I usually work with a student directly… we meet every week to develop it.
—Georgia Perakis
In 15.060 Data, Models and Decisions, we give lectures and do some case studies so students can understand how the material we teach applies in real businesses. To develop a case, I usually work with a student (this student might be an MBA student who took the course already). At the last lecture of the course, I invite students to come find me if they have worked for a company and have a data set with which we can work. Students then come to me with their data and a topic. If they bring to me something that relates to part of the course material I feel has fewer cases, we'll work on developing a case study based on their data. From there, we meet every week to develop it. The student gets an opportunity to collaborate with a faculty member and produce something that contributes to the course.
It's a process. From beginning to finish, it takes about a year to develop a case. We work one semester more intensively, and usually we use the course to roll it out. I begin by discussing what the case is with the student, and then we keep asking questions and iterating, while keeping in mind the parts of the course we want to augment.
We usually work on the business context of the company first, and why this case is relevant. Then we work out aspects of the case to spotlight, how to build the data, the regression, and/or the optimization model, how to understand and make sense of the outputs, and how the company might actually use the model in the case. In general, we might introduce one new case per year.