Knowledge structuring and dynamic modeling via qualitative data analysis: An approach for acceleration using generative AI

Abstract

In recent years, enhancing initial response capabilities during large-scale disasters has become an urgent priority. However, under extreme conditions where unexpected events occur consecutively, there is a limit to what personnel on the ground can handle beyond their prior preparations, regardless of their utmost efforts. Therefore, it is essential to extract the challenges and tacit knowledge faced by practitioners and rapidly codify them into explicit knowledge as decision-support insights for future disasters. While qualitative data analysis is indispensable for uncovering the contexts and underlying rationales of on-site judgments that are difficult to quantify, conventional methods suffer from subjectivity—relying heavily on the analyst—and significant delays due to the immense time required. This study proposes a method to accelerate the qualitative data analysis process utilizing generative AI, and presents an approach to rapidly construct a conceptual model that supports structural understanding based on the lessons learned.