Fingerprints in a Computer Science Course Profile: Early results from the IFNMG DataWarehouse

Abstract

This manuscript reports the implementation of an Educational Data Warehouse (EDW) at the Federal Institute of North of Minas Gerais by using data from the academic system called Cajuí. The logical model of the system is the Fact Constellation with data persisted into the relational DBMS PostgreSQL. After the loading and setting of the EDW, we ran a set of analytic queries regarding courses from the Computer Science bachelor course at the IFNMG campus of Montes Claros for the 2013/2020 timespan. The data analysis indicates: (i) there were no significant differences in the academic performances of students enrolled by either standard entrance or Brazilian SISU exams, (ii) the number of unofficial dropouts reached up to 19% of students, (iii) the 19.51% of students that took a leave of absence and 15.38% of dropout bachelor candidates had completed at least 1/3 of courses from the entire graduation process, (iv) the first-year courses had more failing students than final-year courses, and the average grade of final-year courses was higher than those of other years, and (v) nearly 60% of students had at least one failure in either Algorithms and Data Structures or Calculus courses.