Identifying course trajectories of high achieving engineering students through data analytics

Abstract

In this paper we present findings from a study that compares course trajectories of students who performed well academically and graduated in four years and with those of low achieving student. The goal of this research is to identify factors related to course-taking choices and degree planning that can affect students' academic performance. The data for the study was collected from three majors within an engineering school at a large public university: civil, environmental, and infrastructure engineering (CEIE), computer science (CS), and information technology (INFT). The data includes more than 13,500 records of 360 students. Analysis shows that low performers postponed some courses until the latter end of their program, which delayed consequence courses and their graduation. We also found that low performers enrolled in multiple courses together at the same semester that their counterparts do not usually take concurrently. The methods used in this paper, frequent pattern mining and visualization, help uncover student pathways and trajectories with direct impact for advising prospective and current students. The findings can also be used to improve engineering programs' curriculum.