Learning Analytics and Remote Laboratories: a case study based on the application of LEDA framework and data mining techniques

• Grant number: 21/14908-0
• Support Opportunities: Scholarships in Brazil - Doctorate
• Start date: December 01, 2022
• End date: April 07, 2024
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Marco Antonio Garcia de Carvalho
• Grantee: Carinna Nunes Tulha
• Host Institution: Faculdade de Tecnologia (FT). Universidade Estadual de Campinas (UNICAMP). Limeira , SP, Brazil

Abstract

In order to support digital technologies application into educational context, it is important to understand the relationships between learners and the technologies they use during a practice. Among educational technologies are remote laboratories, tools that provide the manipulation of real experiments through an online platform, available 24/7. That way, remote laboratories overcome restrictions of time and space. To extract information of the big amount of data generated during interactions, it is necessary to use technology-supported representations, in order to apply techniques capable of analyzing and extracting information from data generated from interactions with technologies, enabling learning interventions. A technique called Learning Analytics is based on measuring, collecting, analyzing and reporting student data during practices. Learning Analytics combines data mining techniques to extract information and pedagogical intervention. In this project we propose to develop an educational data mining framework based on Learning Analytics interventions called LEDA (Laboratory Experimentation Data Analysis). The LEDA framework aims to extract information of interaction data with remote laboratories to relate students' interaction behavior with their learning progress. Our approach will apply association rules and clustering techniques using data, including clicks, number of controlled components, and time spent during the activity, among others. The application of association rules allows the survey of conditions and relationships between student's actions and results. By applying clustering techniques, it is possible to organize similar behaviors and define performance on a given activity. We expect to obtain information about the best ways to perform a remote experiment, informing about students' behavior and their engagement during experimentation, in order to maximize learning.