A systematic review on the dimensions of open-source disaster intelligence using GPT

Abstract

Natural and manmade disasters like landslides, floods, earthquake, cyclone, shooting, riots have detrimental effect in precious life, infrastructure, and economy. This study addresses the need for a comprehensive analysis of Generative Pre-Trained Transformers (GPT) in the context of open-source disaster intelligence, a topic where existing literature remains fragmented. Employing a systematic approach, a query scheme incorporating 11 at. keywords was devised, resulting in the acquisition of 53 relevant studies. These studies were meticulously reviewed and synthesized to propose six dimensions of GPT-based open-source disaster intelligence, yielding critical insights into disaster management strategies. Within these 6 dimensions, 24 studies were categorized under “Social Media Analytics for Disaster Response” dimension, 7 on “Disaster Prediction,” 11 on “Disaster Management,” 5 on “Disaster Support Via Technology”, 3 on “Climate Change and Disaster Communication,” and 5 studies were classified under the “General Disaster Analysis” dimension. Leveraging advanced methodologies and machine learning driven tools such as PRISMA, Litmaps, and VOSviewer, this research not only identifies key trends and collaborative efforts but also provides valuable bibliographical insights for researchers and practitioners in the field. For example, the co-citation analysis demonstrated a total of 3703 authors, among whom 51 authors garnered a minimum of 10 citations, leading to the identification of 3 distinct clusters. By addressing a critical research gap and offering a methodologically robust examination, this study contributes significantly to the advancement of knowledge in GPT-based open-source disaster intelligence, facilitating informed decision-making and enhancing disaster response strategies worldwide.