Architecture and Building Enginnering Educational Data Mining. Learning Analytics for detecting academic dropout

The present work is part of a broader research related to the improvement of the teaching methodology, in the undergraduates degree of Architecture and Building Engineering through the application of new technologies. The aim of the proposal is to confirm that changes in the teaching methodology improve the learning experience, in our case using a learning analytics approach. In this case study, we focused in one First Year subject: Descriptive Geometry, which has a high rate of repeating students, as well as an early dropout. We have implemented an educational data mining mixed approach related to the midsemester exams (midterm), and we stablished a relation with the final marks of the subject in two periods with differentiated methodologies, Pre-Bologna (2006-09), and Post-Bologna (2015- 18). The objective of this analysis is to predict what students are closer to leave the course after the midterm results based on the topics examined, so that we can influence and implement new methodologies, technologies and systems to improve these topics.