Mapping the evolution of learning analytics in education: Unveiling trends and insights

Abstract

This study maps research trends in Learning Analytics (LA) through a large-scale bibliometric synthesis covering 2014–2023. Records retrieved from Web of Science and Scopus were de-duplicated and screened with an adapted, transparency-oriented PRISMA workflow, yielding a final corpus of 2245 peer-reviewed journal articles. Using VOSviewer and R, we computed standard indicators (citation, co-citation, co-authorship, bibliographic coupling, and keyword co-occurrence) and thematically examined the 20 most-cited LA papers against the LA reference model. Findings show a marked acceleration of LA publications after 2018; the USA, Australia, Spain, China, and England serve as central collaboration hubs, with Monash University, the Open University, and the University of Edinburgh among the most influential institutions. Keyword structures position “higher education,” “machine learning,” and “educational data mining” as core topics, while AI-enabled and multimodal analytics emerge as growth areas. The analysis also reveals persistent gaps: the “Who” (stakeholders) and, to a lesser extent, the “What” (data environments) dimensions remain underrepresented in seminal work. By integrating two major databases and combining macro-level mapping with a targeted appraisal of landmark studies, this paper offers an up-to-date overview of LA’s evolution and highlights actionable priorities. We conclude that future research should prioritize interdisciplinary and international collaboration, ethical and stakeholder-centric design, and AI-supported, multimodal approaches to enhance impact and generalizability.