Published November 6, 2026 | Version v8

LLM-GeoDis: Subnational Geocoding of Global Disasters Using Large Language Models

  • 1. European Commission, Joint Research Centre, Ispra, 21027, Italy
  • 2. Institute of Health and Society (IRSS), University of Louvain (UCLouvain), Clos Chapelle-aux-champs 30, Woluwe St Lambert, 1200, Brussels, Belgium
  • 3. Institute for Environmental Studies, Vrije Universiteit Amsterdam, Amsterdam, the Netherlands

Description

This repository contains the code and datasets used to produce the LLM-GeoDis dataset, a global database of subnationally geocoded disaster events from the EM-DAT International Disaster Database. The workflow uses a large language model (GPT-4o) to extract and standardize textual disaster location descriptions and match them to administrative units using GADM, OpenStreetMap, and Wikidata.

The resulting dataset provides subnational geocoding for global disaster events recorded in EM-DAT (2000–2024). Each record has been automatically processed to extract location entities and link them to administrative units. The dataset contains 14,215 disaster events across 17,948 unique locations, each associated with GADM administrative levels 1–2. It includes point geometries from Wikidata and OSM as well as harmonized GADM geometries to ensure consistent spatial coverage. Due to its size (~46 GB uncompressed), the full LLM-GeoDis database is distributed across five compressed files.

This Zenodo record contains the following files:

  • LLMGeoDis_part1–5.zip – compressed parts of the main LLM-GeoDis dataset containing the geoparsed and geocoded disaster locations.

  • geoemdat_gaul.gpkg – GeoPackage containing EM-DAT events mapped to FAO GAUL administrative boundaries.

  • pend-gdis-1960-2018-disasterlocations.csv – GDIS dataset of geocoded disaster locations (Rosvold & Buhaug, 2021), used as a benchmark.

  • reliability_db.csv – inputs of the reliability score: availability of each source (GADM, OSM, Wikidata) and agreement between them for each location.

  • input_emdat.csv – EM-DAT input dataset used for the geoparsing workflow.

  • 241204_emdat_archive.xlsx – archived version of EM-DAT used for validation and reproducibility, redistributed under CC-BY-NC-ND (doi:10.14428/DVN/I0LTPH).

  • gdis_disnos.csv – mapping between EM-DAT event identifiers and events in the GDIS disaster dataset used for benchmarking.

  • emdat_geocoding-zenodo.zip – archive of the full code repository used to reproduce the geoparsing, geocoding, and validation workflows.

  • CHANGELOG.md - summarized changes in this version compared to the previous one. 
  • Instructions_LLM-GeoDis.pdf - instructions to run the full code and reproduce the analysis in the manuscript. 

External datasets required to fully reproduce the workflow include GADM 4.1 administrative boundaries and the full GDIS dataset, which must be downloaded separately from their respective sources.

Files

Instructions_LLM-GeoDis.pdf

Files (18.1 GB)

Name Size
md5:e78b986f8234a48de376b8026017d1dc
6.7 MB Download
md5:31431f71b392b444510f11f60c0ba655
3.3 kB Preview Download
md5:e1aab83acd14289afd4e9f0f16e79a0a
33.0 MB Preview Download
md5:41e6a666c7e26323ae6853ef753f960d
135.3 kB Preview Download
md5:79b9e864059c19e6bf76b084a4aba9ee
6.1 GB Download
md5:97384f5817103569d2b1b18de5be4ea4
2.8 MB Preview Download
md5:d18d4816204161701bb4de884a5b9158
177.7 kB Preview Download
md5:0241ebadf2ab2751d40518ea9c420188
2.0 GB Preview Download
md5:c191d5a4bbdef01eefc20f3625b02596
1.9 GB Preview Download
md5:05f6dadb12c6b57e9edec878b48e10cc
2.3 GB Preview Download
md5:3c04ebd80c39cf4f3b9f8bbb011a54c8
2.8 GB Preview Download
md5:f82227f14e15c66d732fe27f041c684c
2.9 GB Preview Download
md5:c6a669466c815a0882deb5b4cc648acb
4.9 MB Preview Download
md5:2b570c081300389231d21435c433b7ac
3.5 MB Preview Download