This article presents a methodology for advancing Learning Analytics by automatically reconstructing the organisational structure of Moodle courses leveraging platform logs and backup data. Using an extensive, pseudonymised dataset gathered from twelve undergraduate courses, we illustrate how student activity records can be enriched with structural information. This methodology integrates xAPI statements with parsed XML data from Moodle course backups to create directed graph representations that capture both hierarchical and sequential relationships among sections and activities. By introducing graph-based structural modelling into Learning Analytics practice, we provide new tools to facilitate replicable and context-aware analytics, enabling deeper insights into the relationship between course structure and learning behaviours, as well as actionable enhancement of digital course design.

Due to the flexibility of online learning courses, students organise and manage their own learning time by deciding when, what, and how to study. Each individual has distinctive learning habits that identify their behaviours and set them apart from others. To explore how students behave over time, in this work we seek to identify adequate time-windows that could be used to investigate the temporal behaviour of students in online learning environments. We first propose a novel perspective to identify various types of sessions based on individual requirements. Most of the works in the literature address this problem by setting an arbitrary session timeout threshold. In this paper we propose an algorithm that helps us in determining the most suitable threshold for the session. Then, based on the identified sessions, we determine time-windows using data-driven methods. To this end, we created a visual tool that assists data scientists and researchers in determining the optimal settings for the session identification and locating suitable time-windows.

Temporal analysis has been demonstrated to be relevant in Learning Analytics research, and capturing time-on-task, i.e., the amount of time spent by students in quality learning, as a proxy to model learning behaviour, predict performance, and avoid drop-out has been the focus of a number of investigations. Nonetheless, most studies do not provide enough information on how their data were prepared for their findings to be easily replicated, even though data pre-processing decisions have an impact on the analysis’ outcomes and can lead to inaccurate predictions. One of the key aspects in the preparation of learning data for temporal analysis is the detection of anomalous values of temporal duration of students’ activities. Most of the works in the literature address this problem without taking into account the fact that different activities can have very different typical execution times. In this paper, we propose a methodology for estimating time-on-task that starts with a well-defined data consolidation and then applies an outlier detection strategy to the data based on a distinct study of each learning activity and its peculiarities. Our real-world data experiments show that the proposed methodology outperforms the current state of the art, providing more accurate time estimations for students’ learning tasks.