Supplementary Data

This repository contains the datasets and prompt resources supporting the analyses reported in the associated study. The materials include processed datasets used to generate figures in the main text and Supplementary Information, literature-derived datasets used in the analytical workflow, and prompt templates used in the large language model–assisted screening framework. The literature data include a cleaned publication database (~5.0 million records), a subset of publications identified as describing real-world disaster events (279,538 records), and a dataset linking publications to specific hazards and countries. Additional files provide prompt templates, human verification records, and model performance metrics. These materials support the transparency and reproducibility of the study.