README.md

---
dataset_info:
  features:
  - name: address_detail_pid
    dtype: string
  - name: street_locality_pid
    dtype: string
  - name: locality_pid
    dtype: string
  - name: building_name
    dtype: string
  - name: lot_number_prefix
    dtype: string
  - name: lot_number
    dtype: string
  - name: lot_number_suffix
    dtype: string
  - name: flat_type
    dtype: string
  - name: flat_number_prefix
    dtype: string
  - name: flat_number
    dtype: float64
  - name: flat_number_suffix
    dtype: string
  - name: level_type
    dtype: string
  - name: level_number_prefix
    dtype: string
  - name: level_number
    dtype: float64
  - name: level_number_suffix
    dtype: string
  - name: number_first_prefix
    dtype: string
  - name: number_first
    dtype: float64
  - name: number_first_suffix
    dtype: string
  - name: number_last_prefix
    dtype: string
  - name: number_last
    dtype: float64
  - name: number_last_suffix
    dtype: string
  - name: street_name
    dtype: string
  - name: street_class_code
    dtype: string
  - name: street_class_type
    dtype: string
  - name: street_type_code
    dtype: string
  - name: street_suffix_code
    dtype: string
  - name: street_suffix_type
    dtype: string
  - name: locality_name
    dtype: string
  - name: state_abbreviation
    dtype: string
  - name: postcode
    dtype: int64
  - name: latitude
    dtype: float64
  - name: longitude
    dtype: float64
  - name: geocode_type
    dtype: string
  - name: confidence
    dtype: int64
  - name: alias_principal
    dtype: string
  - name: primary_secondary
    dtype: string
  - name: legal_parcel_id
    dtype: string
  - name: date_created
    dtype: string
  splits:
  - name: train
    num_bytes: 4657596871
    num_examples: 15357486
  download_size: 1584880457
  dataset_size: 4657596871
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: cc-by-4.0
language:
- en
task_categories:
- text-classification
tags:
- geospatial
- address
- location
- australia
- gnaf
- geoscape
pretty_name: gnaf
size_categories:
- 10M<n<100M
---

# Geoscape Geocoded National Address File (GNAF) 2022

## Dataset overview

Geoscape GNAF is the geocoded address database for Australian businesses and governments. It’s the trusted source of geocoded address data for Australia with over 50 million contributed addresses distilled into 15.4 million G-NAF addresses. It is built and maintained by Geoscape Australia using independently examined and validated government data.

It contains the state, suburb, street, number and coordinate reference or geocode for street addresses in Australia.

Original data source: https://data.gov.au/data/dataset/geocoded-national-address-file-g-naf

G-NAF Product Description: https://docs.geoscape.com.au/projects/gnaf_desc/en/stable/index.html

Transformed into csv using PostgreSQL scripts from: https://github.com/dylanhogg/address-net/tree/master/gnaf_loading

## Columns extracted into this dataset

This dataset contains 15.3 million rows with text and numeric column values extracted from the source GNAF files using PostgreSQL scripts from: https://github.com/dylanhogg/address-net/tree/master/gnaf_loading

## Columns

```
building_name
flat_number
flat_number_prefix
flat_number_suffix
flat_type
latitude
level_number
level_number_prefix
level_number_suffix
level_type
locality_name
longitude
lot_number
lot_number_prefix
lot_number_suffix
number_first
number_first_prefix
number_first_suffix
number_last
number_last_prefix
number_last_suffix
postcode
state_abbreviation
street_name
street_suffix_code
street_type_code
```

## Provenance

### Data source

G-NAF is a complex and large dataset (approximately 5GB unpacked), consisting of multiple tables that will need to be joined prior to use.

This dataset was created from source data from Aug 2022.

https://data.gov.au/data/dataset/geocoded-national-address-file-g-naf

### Collection methodology

Transformed into csv using PostgreSQL scripts from:

https://github.com/dylanhogg/address-net/tree/master/gnaf_loading

## Restrictions

The EULA terms are based on the Creative Commons Attribution 4.0 International license (CC BY 4.0). However, an important restriction relating to the use of the open G-NAF for the sending of mail has been added. The open G-NAF data must not be used for the generation of an address or a compilation of addresses for the sending of mail unless the user has verified that each address to be used for the sending of mail is capable of receiving mail by reference to a secondary source of information.

## Attribution

Incorporates or developed using G-NAF © Geoscape Australia licensed by the Commonwealth of Australia under the Open Geo-coded National Address File (G-NAF) End User Licence Agreement.