<![CDATA[This study aims to map the intellectual structure, publication growth, collaboration patterns, thematic evolution, and research gaps of sentiment analysis in educational technology. The study addresses the lack of an open and integrated bibliometric synthesis that connects productivity, collaboration, topic modelling, and gap detection in this field. A bibliometric and science-mapping approach was applied using OpenAlex-indexed publications from 2013 to 2025. After deduplication and eligibility screening, 977 publications were analysed, while 768 papers with sufficient abstract text were used for Non-negative Matrix Factorisation topic modelling. The analysis included publication trend analysis, country and institutional productivity, co-authorship networks, keyword burst analysis, geographic gap analysis, and platform mention analysis. The results show that annual publications increased from 22 papers in 2013 to 123 papers in 2024, with India and China as the most productive countries. Six thematic clusters were identified: Learning Analytics, Social Media Sentiment, Emotion Recognition, MOOCs and E-learning, Transformers/LLMs, and ML Classifier Ensembles. Learning Analytics was the largest cluster, while Transformers/LLMs showed the fastest recent growth. The novelty of this study lies in its reproducible OpenAlex-based bibliometric framework, which integrates performance analysis, science mapping, thematic evolution, and research gap identification for sentiment analysis in educational technology.]]>
Copyrights © 2026
