Geographic public data,
cross-referenced and ready.

The data preparation is done

The most expensive part of geographic research is usually not the analysis. It is collecting datasets from separate portals, reconciling different geographic identifiers, normalizing conflicting schemas, and QA-testing joins before any actual analysis can begin.

DaedalMap's maintained packs represent that work already completed. Each pack is normalized to a shared geographic schema, QA-gated before release, versioned, and documented. The same loc_id key that identifies a US census tract in the FEMA National Risk Index is the same key used in disaster event data, economic indicators, and climate projections. Cross-domain queries do not require a custom join.

What is in the catalog

Published packs cover:

• Disasters: earthquakes (1M+ events, 2150 BC to present), tropical storms (13,000+ storms, 1842 to present), tsunamis (2,600+ events, 2000 BC to present), volcanoes (11,000+ eruptions, Holocene), wildfires, tornadoes, floods, landslides (45,000+ events since 1760)
• Demographics and development: UN SDG indicators for 200+ countries (2000 to 2023), WHO health indicators for 198 countries, Eurostat for 37 European countries
• Economics: IMF balance of payments for 195 countries, OWID CO2 data for 217 countries (1750 to 2024), global currency exchange rates
• Risk and vulnerability: FEMA National Risk Index at county and tract scale, social vulnerability and community resilience metrics

All packs share the same loc_id geography schema. A query combining earthquake frequency with economic indicators or social vulnerability requires no schema translation.

Research mode

Research mode is a bounded analytical workspace. You build a named corpus from published packs in your account, then load it in the app to ask comparison, synthesis, and trend questions against that specific data.

The corpus boundary is explicit. Research answers from what is in the corpus. If something is missing, it says so. That keeps the reasoning honest and the results auditable.

Bring your own data

The pack schema is public. If you have your own datasets -- survey results, field measurements, custom raster aggregations, institutional data -- they can be normalized to the same loc_id schema and used alongside maintained packs in the same Research workspace.

A climate researcher who brings land surface temperature rasters for a county study can cross-reference them against FEMA risk scores, NLCD impervious surface, and building footprint data in the same query session. The local data travels with the same geographic key as the maintained packs.

Export

Query results export to CSV and Parquet. Both formats are compatible with R, Python, Stata, and QGIS. Pack metadata, source attribution, and coverage scope export alongside the data.

Self-hosted and private data

For research involving sensitive or proprietary data that cannot leave your environment, the same engine runs locally. Install the packs you need, add your own private sources, and run queries with no cloud dependency. 2x to 6x faster than the hosted app in benchmarks. Optional local AI model removes the last external dependency.

Example queries

• How does earthquake frequency in Pacific Rim countries correlate with GDP per capita over the past 30 years?
• Which US census tracts have the highest combined FEMA risk score and social vulnerability?
• How did UN SDG 3 health indicators change across Southeast Asia in the decade following major tsunami events?
• Which Fairfax County block groups have the highest impervious surface fraction and the highest observed land surface temperature?
• What is the historical storm frequency trend for Atlantic basin hurricanes above Category 3 since 1950?