From logs to insights: A comprehensive framework for data-driven learning insights

This study develops a theoretical framework for learning analytics utilizing data from the Moodle Learning Management System (LMS). Despite Moodle's extensive use in educational settings, its potential for learning analytics remains underutilized. This research aims to design a predictive framework for identifying learning difficulties through Moodle's internal analytics, incorporating various data points such as activity completion, attendance logs, social interactions, and learner habits. The study employs a research and development methodology with three main stages: (1) needs analysis and learning component identification, (2) theoretical framework design, and (3) validation through focused group discussions with learning experts. The framework integrates predictive modeling for learning retention, task load analysis, and personalized learning style assessments based on the VARK model. Results demonstrate that the framework effectively uses Moodle's default logs for analyzing learner behavior, although it is limited to online interactions within the LMS. Validation confirms its alignment with Moodle's architecture and online learning theories, with minor adjustments for task load components. The framework offers a scalable solution for institutions managing large student populations and varied learning models, serving as a foundation for early intervention and improved learning outcomes. Future studies could expand the framework's scope to include offline and face-to-face interactions.