Learning analytics

Content sourced from Wikipedia, licensed under CC BY-SA 3.0.

Learning analytics is the practice of measuring, collecting, analyzing, and reporting data about learners and their surroundings to understand and improve learning and the places where it happens. The rise of online education since the 1990s, especially in higher education, has boosted the use of Learning Analytics because more student data can be captured and studied. Data can come from learning management systems (LMS), social media, and other online tools, including what students click, how they navigate, how long they spend on tasks, who they talk to, how information flows, and how concepts develop in discussions. The rapid growth of massive open online courses (MOOCs) provides even more data for researchers.

Definitions of Learning Analytics vary and are still debated. Some early views described LA as using smart data, learner-generated data, and analysis models to find patterns and give advice about learning. Others see LA as a broad framework for guiding analytics services in education—helping with practice, guidance, quality assurance, curriculum design, and teaching effectiveness. A common idea is that LA is a type of Analytics focused on improving learning and education, using data and models to guide decisions.

A widely cited contemporary model (by Gašević, Kovanović, and Joksimović) says Learning Analytics sits at the crossroads of three areas: data science (how to collect, process, analyze, and present data), theory (learning sciences, psychology, sociology, and related fields), and design (learning design, interaction design, and study design). This view emphasizes that computational methods must connect with educational theory and practical design to be effective.

Learning Analytics is often compared with Educational Data Mining (EDM). EDM tends to focus on mining data from educational settings to support administrators and policymakers, while LA tends to focus more on improving learning for students and teachers. Some researchers see these as overlapping areas with shared methods, while others distinguish them by where the questions come from or who the results are meant to help.

The field draws on several disciplinary roots. The main historical sources include four Social Sciences areas that have shaped LA: Social Network Analysis (studying connections and interactions as networks), User Modeling (personalizing systems to individual users), Cognitive Modeling (representing how people think and learn), and Data Mining/E-Learning (extracting patterns from large educational data). These roots helped LA develop its core ideas: mapping how learners interact, personalizing learning experiences, modeling knowledge growth, and using data to improve online education.

Social network analysis looks at how people are connected and how information flows through groups. It considers nodes (people or things) and the links between them. Early work in this area explored how conversations and collaborations form, including the insight that weaker connections can bring in valuable new information. This idea has influenced Learning Analytics by showing that not only strong, obvious relationships matter, but also subtle connections can reveal important learning patterns. Tools and projects in this tradition, such as software for evaluating discussion networks in LMS forums, helped illustrate how learning happens in communities.

User modeling aims to tailor systems to individual needs, goals, and preferences. The idea that computers should treat each learner as a unique person has guided the development of personalized learning experiences and adaptive systems. Hypermedia—nonlinear information networks that combine text, graphics, video, and links—added to this by enabling more personalized navigation and content. Adaptive hypermedia uses a model of a learner’s goals and knowledge to adjust what they see and how they learn.

Cognitive modeling and intelligent tutoring have long explored how to model problem-solving and knowledge. Early ideas about intelligent tutoring systems sought to match instruction to what a learner knows and what they still need to learn. This cognitive focus remains influential in LA, helping researchers design systems that adapt to how students think and learn.

Other theoretical strands include Epistemic Frame Theory, which looks at how people think and act in collaborative learning environments, and Epistemic Network Analysis, which examines connections among ideas and learners. These theories support the goal of making analytics transparent, so educators can understand and validate how insights are reached.

The field has grown into formal education programs. The first dedicated master’s program in Learning Analytics began at Teachers College, Columbia University, in 2015. Since then, other universities have started LA programs as well. Many LA efforts combine ideas from EDM and LA to train researchers and practitioners.

Methods and tools used in Learning Analytics can be applied in various educational contexts. Analytics may track student engagement, performance, and interactions; social network analysis maps connections and discussions; and specialized tools help evaluate learning networks inside LMS platforms. As with any data work, ethics and privacy are important considerations. Questions include what data is collected, how it is used, and how students can control their information. There have been high-profile concerns about trust and data governance, leading to ethics and privacy guidelines. One widely cited approach is a practical eight-point checklist for privacy and ethics in Learning Analytics, designed to help design and implement responsible analytics that benefit students, teachers, and institutions.

Because data collection often happens by default (for example, through cookies in online courses) and students may not always opt out, privacy practices vary by country and system. Open approaches to Learning Analytics aim to improve learning across lifelong learning contexts, focusing on gathering diverse data in ways that respect learners’ rights.

In short, Learning Analytics uses data from online learning to understand how people learn and to improve teaching, learning design, and educational systems. It combines data science with theories of learning and thoughtful design, while continuously addressing ethical and privacy challenges to keep trust and learning outcomes at the forefront.