# Australia geography — 2026-10-24-v1

This is a frozen offline snapshot downloaded on **24 October 2026**. It contains **14,754 accepted settlement points** and **exactly 150 daily answers**. All accepted points can be guessed; there is **no minimum population**. Zero population means unrecorded or unavailable in GeoNames, not proof of no inhabitants.

Scope: Mainland, Tasmania, the ACT, Northern Territory and Jervis Bay Territory. External territories are excluded.

`places.json` uses the Nedle contract: `version`, `snapshotDate`, `places`, `answerIds`, `sources`. Each place has `id` (`geonames:<id>`), `name`, `aliases`, `province`, `municipality`, `lat`, `lon`, `population`, and `eligible`. `eligible` equals membership in `answerIds`. Homonyms retain separate stable IDs. Canonical settlement labels retain the GeoNames point identity; explicit native/common name corrections preserve the former label as an alias. Administrative names never replace settlement labels.

`map.json` contains `land` and `provinces` GeoJSON FeatureCollections in WGS84 longitude/latitude, `sources`, and `centralMeridian`. Natural Earth 1:10m country coastline and internal administrative boundaries are clipped to the game scope, and physical lakes are subtracted. These are overview maps, not navigation or cadastral boundaries. Some countries' Natural Earth first-level administrative layer uses counties/departments/provinces, while the place label uses the country's GeoNames admin1 level. The `provinces` property is a shared UI contract, not a claim that every country calls its subdivisions provinces. Coordinates used for distances are independent settlement points and are never moved to map centroids.

The renderer must unwrap all outline, point and route longitudes around `centralMeridian`. This preserves Chatham and the Aleutian Islands across the antimeridian.

The answer roster is population-led after curation, with regional centres explicitly represented. Current GeoNames PPL, PPLA/PPLA2–PPLA5, PPLC, PPLG, PPLL and PPLF records are eligible as guesses; historical, abandoned, destroyed, grouped-place and PPLX neighbourhood records are excluded. Farms are guess-only. Real districts misclassified by GeoNames as PPL may remain guesses; they are not intentionally selected as daily answers. Nearby identical-name points in the same administrative context merge only within 500 m, preserving aliases; distant namesakes remain distinct.

National registers cross-check names/identity and help remove boroughs, suburbs, composite municipalities and inflated municipal-only labels. A name match validates identity within the stated matching method, not city status, current settlement population or every hamlet. This roster is **not an official current top-150 population league table**. Population values remain the mixed-date GeoNames source values, including known underestimates; they are not overwritten with municipal population totals.

`province` comes from GeoNames admin1. `municipality` uses the most granular available named GeoNames administrative area (admin4/admin3/admin2); where a municipal name is unavailable it uses the known wider district/county/region. This provides useful context without inventing a municipality. The exact hierarchy/fallback is recorded in `place-metadata.json`.

`answers.csv` is the readable fixed roster. `place-metadata.json` includes feature codes, original administrative codes, modification dates and national identity matches. `duplicate-merges.json` records deduplication. `../curation.json` records every manual canonical label, alias, inclusion and exclusion with its reason.

## Attribution and sources

Runtime attribution: **GeoNames (CC BY 4.0) · Made with Natural Earth**.

- [AU.zip](https://download.geonames.org/export/dump/AU.zip): CC BY 4.0; downloaded 2026-10-24. SHA-256 `0ad8e1a7d8ff51b7a3c7b9395ec48a0e3f12a1c60fafb90a74a7c769a9d7941e`.
- [admin1CodesASCII.txt](https://download.geonames.org/export/dump/admin1CodesASCII.txt): CC BY 4.0; downloaded 2026-10-24. SHA-256 `2e4867f5397acb056ea26377601b57af37cb4c81399770225018e9e439e91ad3`.
- [admin2Codes.txt](https://download.geonames.org/export/dump/admin2Codes.txt): CC BY 4.0; downloaded 2026-10-24. SHA-256 `1e7290b22b4e6358e55481cd7e7dd7ae0f8564ad6c67bc2349821ae749c68837`.
- [Australian Bureau of Statistics 2021 UCL allocation names](https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs/edition-3-july-2021-june-2026/access-and-downloads/allocation-files/UCL_SOSR_SOS_2021_AUST.xlsx): CC BY 4.0; downloaded 2026-10-24. SHA-256 `5679d4bf1c62386313631f2690c8f9f0e4e437bdd5a0a040e6862bd4b62a60df`.

Map sources: [Natural Earth 1:10m country boundaries](https://www.naturalearthdata.com/downloads/10m-cultural-vectors/10m-admin-0-countries/), [internal boundaries](https://www.naturalearthdata.com/downloads/10m-cultural-vectors/10m-admin-1-states-provinces/), and [lakes](https://www.naturalearthdata.com/downloads/10m-physical-vectors/10m-lakes/), all public domain. Their exact source hashes are in `map.json`.

The 150-answer roster has 149 recorded national name matches and 1 explicit inclusion exceptions; the full match records are in `../reports/identity-validation.json`. This is an identity check under the country-specific methodology, not a certified current settlement-population ranking. The preparation sources are retained and attributed in `../sources-manifest.json`.

Rebuild instructions, national matching methods, regional-centre inclusions, map limitations and verification are in `../DATA.md`. Run `scripts/prepare-data.py --verify` and `scripts/check-data.py` from `needless-data/`. Future refreshes require a new data version and deliberate review of the daily answer schedule.