Discovering Bottlenecks in a Computer Science Degree through Process Mining techniques

A Higher Education degree is composed by courses which can be organized in areas or modules. Over last years, time invested by students to complete Higher Education degrees has increased. This increment can be caused by the existence of bottlenecks in the courses of academic programs. We aim to carry out an analysis of students' performance to detect courses which represent bottlenecks in the process of completing a degree, because of many student failing compulsory courses. Students' performance can be stored in data sets. Unfortunately, analysis of large data set can lead to scalability problems not being comprehensive applying manual analysis methods because of its extension. We applied Process Mining techniques to overtake these scalability problems. Process Mining is a set of Sequence Analysis techniques to analyze event logs. In this paper, we conducted an analysis of a data set which includes the performance of 612 students applying Process Mining tools. We obtained frequencies of students according time invested to complete a course. We compared these frequencies to detect possible bottlenecks. Finally, some requirements to consider courses as bottlenecks are proposed.