LLM-GeoDis

This dataset provides subnational geocoding for global disaster events recorded in the EM-DAT database (2000–2024). Each record has been automatically processed using a large language model (GPT-4o) to extract and standardize textual location information, cross-referenced with GADM, OpenStreetMap (OSM), and Wikidata. The dataset contains 14,215 disaster events across 17,948 unique locations, each linked to a GADM administrative unit (levels 1–2). It includes point geometries from Wikidata and OSM, and reprojected, gap-filled GADM geometries to ensure complete spatial coverage. The full database (~30 GB) is divided into five parts for easier distribution and handling. All records are linked to their corresponding EM-DAT entries and associated metadata.