Analysing users' access logs in Moodle to improve e learning

Abstract

In this work the UFSC (Federal University of Santa Catarina) and the FGV-RJ (Fundação Getúlio Vargas do Rio de Janeiro) jointly propose the use of a data mining tool to support the analysis of trends, students profiles, as well as to estimate or foresee the usability level of courses being offered, via Moodle, in the Education area. The study carried out by UFSC on the Moodle database allowed a deep understanding of its database, thus making it easier for the Moodle community to execute important tasks, such as the maintenance of the Moodle database, its adaptation following an institutional customization, and, also, a data mart project by the FGV-Online Program to make the necessary analysis possible. In the end of this paper, an example on its applicability is presented, using the association rules technique. Once a data mart oriented to the analysis of the system's usability is developed, various analyses with different objectives can be executed using the database. Some may use the method proposed here or others, including different data mining approaches, such as clustering, neural networks etc. As such, a new contribution is given to the Moodle community.