Desenvolvimento de software para geração de métricas de desempenho e engajamento em AVAs via xAPI

Abstract

The contemporary educational scenario, driven by the expansion of Virtual Learning Environments (VLEs) such as Moodle (a Learning Management System - LMS), demands advanced approaches for evaluating learner performance and engagement. This work explores the integration of technologies and metrics to provide a holistic understanding of learning experiences, both formal and informal, online and offline. The xAPI (Experience API) specification emerges as a modern technical standard for tracking and recording various forms of learning experiences, facilitating the understanding and comparison of results regardless of context or platform. Complementarily, the Learning Record Store (LRS) acts as the centralized repository for these immutable xAPI Statements, ensuring the interoperability and availability of data for analysis. The bridge between Moodle and the xAPI ecosystem is established by the Logstore xAPI, a plugin that converts Moodle log events into standardized xAPI Statements and sends them to an LRS. To deepen the performance evaluation, metrics proposed by Leitão (2017) are addressed, which include Score, Confusion Level, Response Time, Disorder Level, and Comprehension Level, focusing on performance in assessment quizzes. Additionally, the engagement metrics from Li et al. (2021) are presented, such as Views per Content Object, Total Time Spent on Real Visits, Percentage of Course Accessed, Percentage of Feedback Read, and Attempts per Quiz, which quantify student interaction with the LMS. The combination of these methodological perspectives and technologies allows for a granular analysis of learner behavior, providing valuable insights for educators. By transcending the limitations of traditional recording systems, this integration enables more accurate diagnoses, the identification of learning patterns, and the foundation for more personalized and effective pedagogical strategies, promoting academic success in distributed learning environments. The software developed for this final course project is available at https://github.com/utfpr-cm-dacom-bcc/learning-analytics.