Datasets:
File size: 3,145 Bytes
194e8d2 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 | ---
license: cc-by-4.0
pretty_name: World postal codes and administrative units
size_categories:
- 1M<n<10M
task_categories:
- table-question-answering
tags:
- geography
- postal-codes
- zip-codes
- geonames
- administrative-divisions
- gazetteer
configs:
- config_name: postal_codes
data_files: postal_codes-*.parquet
default: true
- config_name: places
data_files: places.parquet
---
# World postal codes and administrative units
**1,836,751 postal codes in 124 countries**, each with its place name and the administrative units it
belongs to (state / province / region, county / district, municipality), plus **172,038 places** (every town with
1,000+ people) with their state and county or district, population, coordinates and matching postal codes.
What it answers:
- **What is the postcode / zip code / PIN code / PLZ of a town?** Look up the `postal_codes` config by `place_name`
and `country_code`.
- **Which state, county, district or province is a town in?** Use `admin1` (state/province/region) and `admin2`
(county/district) in either config.
- **Which places does a postal code cover?** Filter `postal_codes` by `postal_code`.
```python
from datasets import load_dataset
pc = load_dataset("lakerguy16/world-postal-codes", "postal_codes", split="train")
places = load_dataset("lakerguy16/world-postal-codes", "places", split="train")
```
Or with pandas / DuckDB straight from Parquet:
```python
import duckdb
duckdb.sql("SELECT postal_code, place_name, admin1, admin2 FROM 'hf://datasets/lakerguy16/world-postal-codes/postal_codes-*.parquet' "
"WHERE country_code = 'AU' AND place_name = 'Geelong'")
```
**Live lookup for AI agents and scripts:** a free JSON API and MCP server with the same data at
[place-lookup.mnpi.workers.dev](https://place-lookup.mnpi.workers.dev/) (`https://place-lookup.mnpi.workers.dev/mcp`).
## Columns
`postal_codes`: country_code (ISO 3166-1 alpha-2), country, postal_code, place_name, admin1, admin1_code, admin2,
admin2_code, admin3, admin3_code, latitude, longitude, accuracy (GeoNames: 1 = estimated, 4 = geonameid,
6 = centroid of addresses or shape).
`places`: geonameid, name, ascii_name, country_code, country, admin1, admin2, feature_code, population, latitude,
longitude, timezone, postal_codes (codes whose place name matches the town and whose point lies within 25 km of it;
may be empty).
## Coverage
Largest countries by postal-code rows: PT 206,942, AE 178,171, IN 155,570, JP 146,883, MX 144,655, SG 121,154, PE 96,943, ID 81,058, PL 72,899, FR 51,611, RU 43,538, US 41,490, RO 37,915, ES 37,867, TR 36,307, KR 35,583, UA 29,571, GB 27,450, DE 23,296, LT 21,870, AR 20,260, AT 19,225, SE 18,887, AU 18,566, IT 18,415. Countries without a national postal-code file in GeoNames are only in
`places`.
## Source, license and updates
Built from [GeoNames](https://www.geonames.org/) postal code and gazetteer exports (CC BY 4.0). This dataset is CC
BY 4.0; please credit GeoNames. Rebuilt monthly; last build 2026-11-19 UTC. Postal-code coverage and accuracy vary by
country (see GeoNames' notes); for legal or delivery use, check the national postal operator.
|