Key points are not available for this paper at this time.
Higher education is vital to develop human potential and driving economic advancement. Nonetheless, the widespread occurrence of student dropouts raises significant concerns. Educational institutions must delve into the underlying causes of these dropouts to devise strategies, interventions, resource allocation, and comprehensive support mechanisms. Process Mining (PM) fits well to extract insights from recorded data, providing a comprehension of the trajectories and outcomes among programs, courses, and students. However, PM struggles to identify and represent multiple semantic levels inside a given data instance. This paper introduces the notion of color-map to associate different context representations of a data record to a label that differentiates it from other possible representations in the same data instance. This tends to improve on the PM potential to represent knowledge. A case study applies the method to a comprehensive database from a Brazilian public university, composed of 437,690 events spanning eight distinct programs and students enrolled through the Unified Selection System (SISU). The results show that the method can provide students and educators with indicators to consider when interfering in the process.
Southier et al. (Mon,) studied this question.