Videolesson: Data sources in learning analytics

Watch the video lecture on different types of data sources used in learning analytics, including LMS data, data from learning design tools, multimodal data, and other relevant sources.

 Transcript

	Welcome. In this video lecture, we will explore the data sources used in learning analytics. You may recall from the introductory video lecture that the data collection phase is the first stage in the learning analytics cycle, and it lays the foundation for the subsequent phases. Learning analytics uses different types of data. Broadly, they can be grouped into several main categories: data from learning management system, LMS, student data from other information systems, data from student self-reporting, multimodal data and data from external systems. When we analyze how the learning and teaching process is designed, data from learning design tools such as the Balanced Design Planning BDP tool are essential. Data from learning management systems are the most commonly used in learning analytics. Various types of data can be collected from systems such as Moodle or Blackboard, for example, student activity records can be collected. These logs every action a student takes within the LMS known as log data. Data can also be collected on the time spent on specific activities or on the use of learning resources. Such data capture patterns in student behavior, engagement, and navigation through a course. In addition to such data, which show student activity in the system, assessment data are very important. Specifically student scores and grades obtained in assessment activities can provide valuable insights related to the achievement of learning outcomes. Finally, student actual work, such as their essays can also be analyzed. Student data are also stored in other information systems such as the Higher Education Information System (ISVU) in Croatia. Such systems may provide demographic data or academic history data, which can be useful for understanding issues related to inclusion. They can also be combined with data from other sources like Moodle to gain deeper insight. Learning analytics can also use data from direct student self-reporting, such as students responses to questionnaires, their self-assessments and reflections. These data may (but do not have to be) collected in an LMS, however, they provide a different kind of insight, for example, into students' needs and expectations or their satisfaction with the available learning analytics. In recent years, increasing attention has been given to multimodal learning analytics, which uses data from multiple sources to capture cognitive, emotional and physical activity during learning. For example, this may involve video and audio recordings to capture verbal discussions, eye tracking data to monitor students' focus and attention or data collected by physiological sensors such as heart rate data. Although promising multimodal analytics are used mainly in experimental and advanced research settings. Data related to learning can also be collected in external systems, such as platforms for open educational resources or library systems. These sources can complement the previously mentioned data to provide the broader picture or learning behavior tools such as the balanced design planning BDP tool allow learning analytics to include not only data related to the implementation of the learning process, but also the learning design itself. For example, the BDP tool provides teachers with real-time access to detailed analytics of their designs, helping them plan in line with fundamental pedagogical principles, especially constructive alignment. In addition, it supports planning in accordance with innovative pedagogies and appropriate student workload. By combining different data sources and types of data, we can gain a more holistic view of learning. For example, log data from an LMS can show time spent on activities, but only when combined with assessment results can we understand the actual quality of learning. Likewise, assessment data can be interpreted more meaningfully if we also consider the course design and where the constructive alignment has been achieved. Of course, it is essential to emphasize that the use of data for learning analytics purposes always requires compliance with ethical principles and legal and institutional regulations related to data protection and privacy. Not all data sources are always available in all institutions and educational contexts. Their use  depends on technical, organizational, and legal possibilities. In addition, data quality and representativeness are crucial for the accuracy of analysis. If data are incomplete, biased, or misinterpreted, the conclusions may be incorrect and may even harm the learning process. A critical understanding of data is not only a technical issue, it is essential for using learning analytics responsibly, ethically, and in the service of better learning.