Analyzing course sequencing to find optimal student pathways to graduation

In this example, we’ll work with a dataset representing a medical program to understand pathways to the successful completion of a medical degree. In a real medical program, we’d likely include all courses and potentially other factors, such as clinical experiences and research projects required for graduation. However, to run a simple example, we’ll assume this data mining has already been done to identify courses related to graduation outcomes.

Introduction to a dataset

Let’s imagine a medical program with many courses leading to a final licensing exam. Some courses aren’t emphasized by the final licensing exam very much (but are still important to study before entering the field). A handful of courses, though, do show up regularly on the licensing exam, and some build on prior important licensing courses. Let’s suppose human anatomy, cellular biology, pathology...