E-learning as a new educational paradigm

In the last two decades, the digital transformation of education has brought about a series of profound changes in the way the teaching process is organized, implemented and evaluated. The development of e-learning, which was initially developed as a complement to distance education, and is now increasingly used as an integrated part of higher education, has opened up space for the systematic collection and analysis of large amounts of data on student behavior and achievement. It is precisely this data availability that has created the foundations for the development of learning analytics, a new interdisciplinary approach that views education not only through a pedagogical and didactic prism, but also through the possibilities offered by computer science and the processing of large data sets.

Learning analytics is defined as the process of collecting, measuring, analyzing and interpreting data about students and their educational contexts with the aim of understanding and optimizing the learning process and making informed pedagogical decisions. Although it shares its roots with the evaluation of the educational process, learning analytics focuses on the processing of real, most often digitally collected traces of user behavior, such as activities in the e-learning system, time spent with certain content, quiz results, interactions on forums, or navigation methods in teaching material.

In the context of higher education, the application of learning analytics can encompass multiple levels. At the student level, it enables the provision of personalized feedback and guidance to students based on their learning needs, habits, and achievements. At the teacher level, it provides valuable insights into the effectiveness of teaching materials and methods, uncovers patterns in learning that may indicate the risk of dropout or low achievement, and helps design more effective curricula. At the institutional level, analytics can serve strategic management, program evaluation, and data-driven educational policymaking.

Theoretically speaking, learning analytics relies on several disciplines: pedagogy serves to understand pedagogical concepts and goals, computing and information sciences develop tools for data processing and visualization, while psychology contributes to the interpretation of behavior and motivation. Such interdisciplinarity requires careful interpretation of the results because quantitative data cannot fully cover all dimensions of learning.

It is important to emphasize that the effective application of learning analytics is not limited to the technical implementation of tools, but requires pedagogically thoughtful design of metrics and indicators, clear ethical regulation of data collection and use, and active involvement of teachers in the interpretation of results. Also, the integration of analytics into the higher education context presupposes digital literacy of teaching staff, institutional support, and continuous education of all stakeholders in the education system.

Examples

In higher education institutions using Moodle LMS, learning analytics can be applied in various ways to improve the quality of teaching and provide timely support to students. Using Moodle Learning Analytics , it is possible to conduct various analyses and predictions based on models that automatically process data on student activities and performance. These models use predictive analytics algorithms and can help teachers in early recognition of patterns that indicate risk, progress or engagement, e.g.:

Prediction of students at risk of dropping out

This model analyzes student behavior patterns in a course, including the number of sign-ups, activities completed, forum participation, and quiz scores. Based on the data collected, the system predicts which students are showing signs of low engagement or are at risk of dropping out. The instructor is notified and can take timely interventions, such as offering additional tutoring or personalized assignments.

Prediction of probability of successful completion of upcoming activities due

Moodle can predict the likelihood that a student will complete assigned activities on time. The system analyzes past assignment submission patterns, time spent on the platform, and participation in previous activities. Based on this data, it automatically warns students if there is a risk of tardiness, thus encouraging the development of self-regulation and responsibility in learning.

These models are available in the built-in Moodle Analytics API and can be activated at the e-course or system-wide level. Their purpose is to help teachers and students make data-driven decisions and develop proactive learning support strategies.