HERA — the Human Entity Reference Atlas — discovers, normalizes, and packages public census and human-geography data for every country on Earth. Aligned to DGIWG SEL and CCL, cross-walked to Overture Maps via stable GERS IDs. Official sources first. Every field preserved.
Every national statistics office publishes differently — APIs, PDFs, web tables, bulk downloads, restricted access. HERA handles all of it, and preserves every field along the way.
Primary data pulled directly from NSOs — Census Bureau, StatCan, ABS, ONS, IBGE, INEGI, INSEE, GUS, e-Stat, and more. Secondary sources (World Bank, WorldPop, DHS, MICS, HDX) supplement where needed.
If a country publishes at six admin levels, we create six tables — not one. 77 countries ingested at admin depth 3 or deeper: tract, ward, gmina, neighborhood.
Every field from every source is captured in the DatasetFieldInventory. If the DGIWG schema can't hold a field, an extension table does. Nothing is silently dropped.
Spatial/demographic data maps to DGIWG SEL; cultural/linguistic data to CCL. Admin boundaries and places are built to cross-walk to Overture Maps via stable GERS IDs (US counties crosswalked today, expanding with GERS coverage). ISO 19115 XML ships with every package.
CSV, Parquet, GeoJSON, Excel, and ISO 19115 metadata XML — every country package ships in every format, with provenance, QA reports, and licensing baked in.
3,311 US counties scored for 15 business types: hotels, casinos, dispensaries, restaurants, warehouses, and more. Weighted by tax environment, regulatory status, and pending legislation.
From population and housing to religion, language, and migration, through to socio-political dynamics and vulnerability & risk — HERA captures every thematic layer that matters for defense, development, and commercial decisions.
Age, sex, race, household composition, fertility, mortality.
Tenure, vacancy, rent burden, housing units, affordability.
Religious affiliation, denominations, practice rates, ecclesiastical bodies.
ISO 639-3, genealogy, endangerment, speakers, literacy.
Ethnic groups, tribal affiliations, cultural practices, values.
Stocks, flows, remittances, refugees, IDPs, asylum seekers.
Employment, income, sector breakdown, wages, informality.
DHS indicators, UNESCO enrollment, literacy, attainment.
Ethnic power access and exclusion, governance and criminal-actor scoring, institutional trust, government type.
Hazard exposure, socioeconomic vulnerability, coping capacity, displacement, deprivation indices.
Interactive SEL/CCL world map. Drill into any country, inspect ingested admin levels, datasets, field inventory, and download packaged exports.
Rank every US county for hotels, casinos, dispensaries, restaurants, manufacturing, warehousing, and more. Tax friendliness, regulatory gates, and live legislation.
~50 commercial variables per county: vacancy, rent burden, MSA indicators, HUD income limits, BLS sector employment, FBI crime detail, RPP, and a 0–100 composite score.
District-level search, filters, and county rollups. Enrollment, demographics, and funding tied back to the commercial and site-selection layers.
Type a county name or FIPS and get a single consolidated brief — demographics, commercial operating environment, tax position, schools, and safety — assembled from the same layers as the rest of the platform.
Official NSO APIs wherever possible, bulk download and PDF extraction where needed, and secondary sources where primary data is restricted.
The question an evaluator asks first — which themes, which administrative depth, which census vintage, for which country — is answered from the production database and published in full. Not a claim about coverage: the coverage itself, regenerated from the corpus so it cannot drift from what ships.
Every country HERA holds data for, with depth held, themes, census vintage and SEL/CCL field counts. Depth held is coverage; depth declared is not — both are shown, and the gap between them is named as a priced enhancement rather than folded into a coverage number.
Open the matrix →Every upstream source carries a change-detection rule keyed to its own publication schedule. What has been checked, what changed, and what could not be reached is reported on one page — including, plainly, when a source is not being watched at all.
Monitoring status →Every dataset keeps the publisher's own bytes, with the retrieval URL, date, size and SHA-256 recorded. Delivery is bulk download (CSV, Parquet, GeoJSON, GeoPackage, Excel), a REST API, and the hosted platform; ISO 19115 metadata ships with every package.
API reference →Counts are country totals from the coverage matrix, generated from the production database. The second column is the part most data vendors leave you to discover after signing.
Commercial terms are quoted, not published. Data-rights terms for internal analysis, derivative products, redistribution and integration into Government systems — and how thematic layers divide between a base and an add-on licence — depend on the scope being licensed. Ask for the terms for your scope →
Every figure below is generated from the production database and linked to its evidence. Where the answer is “no” or “quoted”, it says so.
All of them, per country, in the coverage matrix — themes held, administrative depth, census vintage and SEL/CCL field counts for each of 278 countries. Downloadable as CSV or JSON.
A country licence includes every theme HERA holds for that country — there is no per-theme add-on tier. Specialist derived products, such as the USA County Commercial profile and the Russia Federal Subjects set, are separate SKUs.
278 countries, with the depth actually ingested named per country in the matrix. On update cycles: every source carries a change-detection rule on its own publication schedule, and what has been checked, what changed and what could not be reached is published on the monitoring page — including, plainly, when a source is not being watched.
No. HERA holds migration stocks, flows, displacement and refugee data for 226 countries. It does not hold device-derived or telemetry-based movement data.
Roads, power, water and ports are held for 3 of 278 countries. Broader coverage is a priced enhancement rather than something a contract unlocks.
HERA is a transformation pipeline, not only a corpus. The path that produced everything above — acquire, normalise, map to DGIWG SEL or CCL, align to the publisher’s own administrative hierarchy, preserve every field — is source-agnostic. It already handles APIs, bulk downloads, web tables and PDF-only publications, which is what national statistical offices actually publish.
So a dataset HERA does not hold today, or one you bring us, goes through the same pipeline and comes out in the same shape: global, or scoped to the countries and administrative levels you need, in the same formats, with the same ISO 19115 metadata and the same field-level provenance. Nothing is silently dropped — where a field will not fit the standard it is preserved in an extension table, and where a source’s licence forbids redistribution the field is withheld and recorded as withheld.
That is the answer to the infrastructure gap above, and to any other gap: the limit is licensing and lead time, not capability. Tell us the dataset →
Ethnicity 217, religion 213, tribal 148, social structure 201.
Attitudinal survey data — World Values Survey, the regional barometers, Ethnic Power Relations — is not part of the product. Those sources restrict commercial redistribution, so we hold them in a separate store and can supply them to organisations that hold their own licence. Ask us if that applies to you.
Standard: the INFORM Risk Index — hazard, vulnerability and coping capacity — for 191 countries, shipped with every package that covers them. Custom: bespoke indices or re-weightings are scoped and quoted; nothing bespoke is implied by the standard licence.
Published in Terms of Service §4: a non-exclusive, non-transferable licence to use the data products for your internal and commercial purposes. Underlying datasets keep their own source licences, attribution must be preserved, and redistribution is bounded by the upstream licence — where a source forbids it, those fields are withheld and recorded as withheld. Terms for integration into Government systems are scoped per engagement.
Bulk download in CSV, Parquet, GeoJSON, GeoPackage, Excel and Shapefile; a
REST API; and this hosted platform. ISO 19115 metadata ships with every
package. Enterprise GIS integration is live: an
OGC API — Features endpoint at /ogc
that ArcGIS Pro and QGIS add as a service rather than a file, so the layer
stays current instead of going stale the moment it is downloaded. It serves
administrative boundaries at every level HERA holds geometry for —
national, regional and district — read-only, in CRS84.
Thematic attributes are not on the service: census, population
and CCL fields ship as files and join on the identifiers it returns.
Country licences are priced by the administrative depth actually held, with an annual licence covering refreshes; a full-database option is available. Figures are quoted against the scope you need rather than published here. Ask for a quote →
Whether you contribute to Overture Maps, build products on top of it, or license its data downstream, you get places, buildings, and administrative divisions — but no context about the people who live inside them. HERA is that context layer. We pull census, demographic, cultural, and linguistic data directly from national statistics offices across 278 live countries (with 248 countries and territories configured and ingesting), built to join to Overture via stable GERS (Global Entity Reference System) identifiers — the persistent ULIDs Overture assigns to every division. US counties are GERS-crosswalked today; the join key is designed to extend to every admin unit as GERS coverage grows.
If you're an Overture contributor, HERA is the enrichment layer your customers keep asking for. If you resell or redistribute Overture data, HERA is the value-add that lets you charge a premium. If you're building products on Overture — mapping, LBS, geo-AI, retail siting, logistics, risk — HERA is the fastest way to turn empty polygons into decision-ready rows. Building the 248-country-and-territory ingestion pipeline in-house takes years of engineering across API, bulk, scrape, and PDF extraction paths. We've built the hard part. Plug us in.
Population, age, sex, households, housing, income, employment, education — pulled directly from US Census, StatCan, ABS, ONS, IBGE, INSEE, INEGI, GUS, e-Stat, and 30+ more. Joined to Overture division GERS IDs at every admin level the country publishes.
Overture stops at the physical world. HERA's CCL layer adds the cultural and linguistic dimensions: religious affiliation, 11,161 languages with genealogy and endangerment, ethnic groups (D-PLACE / EPR), refugee/IDP flows, and cultural values (WVS / regional barometers). A category Overture simply doesn't carry.
-- Enrich any Overture division with HERA census + CCL in a single join SELECT o.id -- Overture GERS ID (stable across releases), o.names.primary -- division name, x.country_code, x.admin_level, x.local_geoid, t.population_2020, t.median_household_income, t.total_housing_units, t.total_vacancy_rate_pct, t.pct_broadband, t.pct_rent_burdened, t.avg_commute_minutes FROM overture.divisions o JOIN hera.gers_crosswalk x ON x.gers_id = o.id JOIN hera.census_tables t ON t.id = x.census_table_id WHERE o.country = 'US' AND o.subtype = 'county'; -- gers_crosswalk: 3,134 of 3,202 USA admin rows (national + state + county) carry real Overture GERS IDs (97.9%) -- Remaining 68 are categorized placeholders (AK reorganizations, VA BEA pseudo-FIPS, name-match gaps) — see /docs/overture -- Crosswalk pegged to Overture release 2026-08-19.0; 3,243 real GERS IDs across 6 countries and growing
Two honest numbers: 97.9% of all the admin units HERA holds carry a stable Overture GERS ID today; and where Overture models a division at our granularity, HERA matches 97.9% of them. The gap is where Overture models no division at that depth (e.g. Brazil's municipalities) — we hold the authoritative, commercial-safe boundaries for those and are contributing them back to the GERS registry. Read the crosswalk methodology →
Every admin unit HERA holds is matched against Overture — where Overture models the division, it carries a stable GERS join key: a persistent ULID that doesn't drift across releases.
Per-field source, license, and provenance in the DatasetFieldInventory. HERA's own attribution data is licensed for commercial redistribution and carries no share-alike obligation — it enriches your pipeline without changing the terms your pipeline ships under.
Every country, every admin level, every field preserved. API, bulk download, web-scrape, and PDF ingestion — we already handle all of it. You get a single normalized feed.
Open the world map or jump straight into Site Selection. Everything runs locally against the catalog DB.