Mapping the Evolution of Sentiment Analysis in Disaster Management: A Bibliometric Analysis

<p dir="ltr">In order to conduct a thorough and transparent bibliometric evaluation of the academic works related to sentiment analysis in crisis settings, the present study employs a quantitative bibliometric methodology based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework. Bibliometrics is an application of mathematical and statistical tools that helps understand the intellectual structure, trajectories, and collaborations in a certain field of research.

The findings outline the paradigm shifts occurring in sentiment analysis and disaster research between 2011 and 2026. This evolution has shifted the domain from early static lexicons and basic supervised classifiers to multi-modal deep learning architectures, Large Language Models (LLMs), and dynamic noise-filtering mechanisms capable of converting real-time social streams into actionable situational awareness. Increasingly, this research trajectory is propelled by urgent concerns surrounding climate-induced disasters, where social media serves as a vital digital gauge of public risk perception, panic, and community needs for emergency response planning.