A multidimensional data model for the analysis of learning management systems under different perspectives

Abstract

The decision-making process in the educational context has been widely investigated as an effective mechanism to support educators and administrators in distance and blended learning. Pattern discovery in data produced by Learning Management Systems (LMS) elicits important information about the educational process, especially when considering correlations between stakeholders interacting in different perspectives of the LMS like grades, access, forum or message communications. In this paper, we propose a data and metadata model for Learning Analytics applications using data from LMS. This model is LMS independent and supports Business Intelligence and Data Mining applications for decision-making in the educational context. Our model considers the use of historical data from LMS treated under three perspectives that can be analyzed individually or correlated. The model also allows analysis considering different stakeholders (student, educator, etc.) and levels of granularity (institution, department, course, discipline, etc.). In this sense, the proposed model allows consistent and correlated analysis of data from LMSs. As a result, our model effective collaborates with the educational process through an efficient decision-making.