Videolesson: Introduction to learning analytics

Watch the video lesson with an overview of basic concepts and topics related to learning analytics, including the definition, significance, and key areas of application. The concept of human-centred learning analytics is also explained as a connection to learning design and its place in the cycle of continuous improvement.

Transcript

	Welcome to the introductory video lecture on learning analytics. In this lecture, we will introduce the basic concepts and explain the importance and applications of learning analytics, as well as its key challenges. According to the definition of the Society for Learning Analytics Research (SOLAR) key international organization, promoting research and practice in learning analytics. Learning analytics refers to the collection and communication interpretation, and communication of data about learners and their learning that provides theoretically relevant and actionable insights to enhance learning and teaching. It is a relatively new but rapidly developing field of research and practice that connects education, data science, mathematics, and psychology, with artificial intelligence also playing an important role. Although data have long been used to improve learning and teaching, learning analytics brings innovation in terms of using new digital data in learning activities, and techniques from data science and artificial intelligence. As a result, learning analytics includes a variety of methods ranging from standard statistical techniques to more complex ones based on artificial intelligence and machine learning. The goal is to enable understanding of data related to learning and teaching, and to ssupport decision-making learning analytics draws on various data sources, particularly data from learning management systems. For example, logs of student activity and grades from learning management Systems such as Moodle are analysed. Learning analytics also uses multimodal data from different sources in both digital and physical environments. For instance, video recordings and data collected via sensors are used. Learning analytics is becoming increasingly important in creating meaningful learning design, especially in ensuring constructive alignment between learning outcomes, learning and teaching activities, and assessment. There is also learning design analytics, which uses data on planned learning and teaching activities based on data learning analytics provides deeper insight into learning processes and gaps in understanding, enabling more precise monitoring and evaluation of student progress. It thus provides teachers with the basis for delivering constructive feedback to students, creating personalized approaches and implementing timely interventions for students experiencing difficulties or who are at risk of dropping out. It also enables the evaluation of pedagogical soundness and meaningfulness of learning design and its improvement. For students,learning analytics is important in fostering self-regulated learning. Learning analytics varies in complexity, and we can distinguish between descriptive, diagnostic, predictive, and prescriptive analytics. Descriptive learning analytics uses data to identify and understand trends and the current state. For example, it can analyze how much time students spend on specific activities in the learning management system. The percentage of students that watched a particular video. Or the average exam performance. Diagnostic learning analytics is more complex and aims to provide a better understanding of why something happened. For instance, it includes analysis to understand why students achieved poorer results. Predictive learning analytics examines past trends to predict future outcomes and identify areas for improvement. For example, it can analyze which types of learning materials are good predictors of success in the final exam, prescriptive learning analytics provide specific recommendations such as personalized learning advice. Predictive and prescriptive analytics most often rely on advanced methods such as machine learning for prediction and recommendation. While machine learning methods and algorithms such as random forests are already commonplace in learning analytics, generative models based on large language models such as ChatGPT are raising new issues in the field of learning analytics. One aspect concerns conducting analyses using generative AI tools as research instruments, which can make learning analytics easier and more accessible to a wider range of users by bypassing complex tools and programming languages such as R. The second aspect is learning analytics that examines the application of artificial intelligence in learning and teaching as a research subject. For example, the effectiveness of generative AI in assessing student essays is analyzed. Both aspects open numerous questions related to trustworthiness and ethics. The implementation of learning analytics is part of a cyclical process. This process includes data collection and analysis, reporting results, and ends with specific interventions. For example, interventions may include revising learning design, or creating additional materials for students who need extra support through interventions. Learning analytics gains practical application in which data informed decisions are made, actions are taken and improvements are implemented. Therefore, closing the loop is essential for learning analytics to have a real impact. The results of analytics are often presented via dashboards as visual interfaces. They enable students, teachers, and other stakeholders such as decision-makers to view data and analytics results in real time. For example, students can see detailed information such as comparisons with the group average or tracking their activity during the semester. This helps them more easily identify areas that require additional attention or intervention. Learning analytics is closely related to the principles of human-centered design. In this sense, human-centered learning analytics means involving relevant stakeholders in the development of learning analytics and associated tools. For instance, students and teachers actively participate in the design of dashboards, express their needs, and provide feedback, ensuring that the tools truly meet user needs and are effective in practice. Although it is clear and widely they accepted that learning analytics has great potential to improve learning and teaching, there are also significant challenges related to its wider adoption. Some are related to institutional strategies, organizational structures, data quality and access or competencies. Important challenges are related to the trustworthiness of learning analytics and include issues such as ethics, privacy, transparency, algorithmic fairness, and the use of artificial intelligence, including generative models. We will discuss these issues in more detail in one of the following units. There are several sister fields closely related to learning analytics. They have some overlap and use similar tools, but pursue different goals. It is important to mention educational data mining, which also applies machine learning, data mining algorithms and statistical methods to educational data, but is generally considered neutral in terms of pedagogy and educational theory. However, while learning analytics uses data to improve learning and teaching, educational data, mining focuses on algorithms for pattern discovery, with an emphasis on research rather than on application in education. Another related field is academic analytics, which has a different focus and helps institutions make strategic decisions about resource management, finances and employment policies. In conclusion, learning analytics is an important element of the future of education, providing a foundation for evidence-based decision-making to improve the quality of learning and teaching.