ABS-derived boundary files

The purpose of this repository is to provide a convenient method for downloading the various boundary files from the Australian Bureau of Statistics (ABS) and converting them into a consistent format. Some benefits of this code include:

• A single place to get all the ABS boundary files across 2011, 2016 and 2021

• Automatically downloads all necessary files

• Consistently formatted column headings across the years (SA1_MAIN11, SA1_MAINCODE_2016, SA1_CODE_2021 are now all now sa1_code)

• Includes population-weighted centroids for each boundary type that are guaranteed to lie within the original region’s geometry (i.e., no centroids in the ocean)

• Manually extracting the population and dwelling counts from ABS excel spreadsheets is usually an error-prone manual task. This code automatically converts them into clean CSVs

• Some suburbs can lie in multiple LGAs. This code will store the largest two for each suburb

• Produces a single sqlite file for each census year, which can be used across projects

• Includes code that can filter the boundaries down to just the cities or states required, optionally including a user-specified buffer

Output files are available here.

Quick setup

This assumes you have R installed

# install any missing R packages
Rscript installPackages.R
# download the raw boundary files from the ABS
Rscript downloadData.R --year='2021'
# convert the data into a consistent output format
Rscript convertData.R --year='2021'

---

downloadData.R

Synopsis

Rscript downloadData.R [--year]

Description

downloadData.R downloads data for a given year from the abs website. It also converts meshblock count spreadsheets into plain csv files and extracts geopackage files from zips. If a file is already downloaded, then it will not be downloaded again.

--year

Specify the year desired. Currently 2011, 2016 and 2021 are supported.

Example

Rscript downloadData.R --year='2011'
Rscript downloadData.R --year='2016'
Rscript downloadData.R --year='2021'

---

convertData.R

Synopsis

Rscript convertData.R [--year]

Prerequisites

• downloadData.R

Description

convertData.R converts the downloaded data into a single sqlite file. The following tables are created:

• australia
• state
• cities
• lga (local government areas)
• ssc (state suburbs/suburbs and localities)
• poa (postcodes)
• sa4 (statistical area level 4)
• sa3 (statistical area level 3)
• sa2 (statistical area level 2)
• sa1 (statistical area level 1)
• mb (meshblocks)
• state_centroids
• cities_centroids
• lga_centroids
• ssc_centroids
• poa_centroids
• sa4_centroids
• sa3_centroids
• sa2_centroids
• sa1_centroids
• mb_centroids

--year

Specify the year desired. Currently 2011, 2016 and 2021 are supported. Note: 2011 lacks an Australia shapefile so the one from 2016 is used instead. downloadData.R will have to be run for both years before 2011 can be converted.

Example

Rscript convertData.R --year='2011'
Rscript convertData.R --year='2016'
Rscript convertData.R --year='2021'

---

filterToStudyRegion.R

Synopsis

Rscript filterToStudyRegion.R [--year] [--cities] [--states] [--buffer] [--epsg] [--filename]

Prerequisites

• convertData.R

Description

filterToStudyRegion.R filters the converted data to the desired study region

--year

Specify the year desired. Currently 2011, 2016 and 2021 are supported.

--cities

Comma separated list of cities to filter to. Not mandatory, but cannot be used with --states. Can list a single city.

--states

Comma separated list of states to filter to. Not mandatory, but cannot be used with --cities. Can list a single state.

--buffer

Specify the buffer distance in metres around the desired cities or states to be included. Not mandatory.

--epsg

Specify the EPSG (i.e., coordinate reference system to transform the data into). Not mandatory, but if nothing is specified, the source dataset's EPSG will be used.

--format

Specify the file format for your output dataset. Defaults to 'sqlite' but geopackage 'gpkg' is also possible

--filename

Specify the filename for your output dataset.

Example

# A Victoria study region in sqlite format
Rscript filterToStudyRegion.R \
  --year='2016' \
  --states='Victoria' \
  --epsg='7899' \
  --format='sqlite' \
  --filename='Victoria'

# Melbourne and Sydney plus a 10km buffer in geopackage format
Rscript filterToStudyRegion.R \
  --year='2021' \
  --cities='Greater Melbourne, Greater Sydney' \
  --buffer='10000' \
  --epsg='7845' \
  --format='gpkg' \
  --filename='MelbourneAndSydney'

---

Useful Coordinate Reference Systems

Geographic coordinate reference systems use latitude and longitude, measuring distance in degrees whereas projected ones use easting and northing, measuring distance in metres. Any geospatial analysis that involves distances will generally use a projected coordinate reference system.

EPSG	Name	Extent	Notes
7845	GDA2020 / GA LCC	Australia	Best projected CRS for Australia-wide distance calculations
4326	WGS 84	World	Default geographic CRS for Australia. Useful for webmaps
7842	GDA2020	Australia	Latest geographic CRS for Australia
7899	GDA2020 / Vicgrid	Victoria	Latest projected CRS for Victoria
8058	GDA2020 / NSW Lambert	NSW	Latest projected CRS for NSW
8059	GDA2020 / SA Lambert	SA	Latest projected CRS for SA

MGA Zones:

MGA Zones are more accurate than the state or country-wide projections, but cover a much smaller extent. There are seven MGA zones covering mainland Australia and they are typically used for projects where the region of interest is city-scale. It's generally recommend to use the new MGA zones.

MGA Zone	Older EPSG	Older Name	New EPSG	New Name
50	28350	GDA94 / MGA zone 50	7850	GDA2020 / MGA zone 50
51	28351	GDA94 / MGA zone 51	7851	GDA2020 / MGA zone 51
52	28352	GDA94 / MGA zone 52	7852	GDA2020 / MGA zone 52
53	28353	GDA94 / MGA zone 53	7853	GDA2020 / MGA zone 53
54	28354	GDA94 / MGA zone 54	7854	GDA2020 / MGA zone 54
55	28355	GDA94 / MGA zone 55	7855	GDA2020 / MGA zone 55
56	28356	GDA94 / MGA zone 56	7856	GDA2020 / MGA zone 56

Cities and towns in each MGA Zone:

Capital cities are in bold

Zone 50	Zone 51	Zone 52	Zone 53	Zone 54	Zone 55	Zone 56
Albany	Broome	Greater Darwin	Alice Springs	Ballarat	Airlie Beach - Cannonvale	Armidale
Bunbury	Esperance		Port Augusta	Broken Hill	Albury - Wodonga	Ballina
Busselton	Kalgoorlie - Boulder		Port Lincoln	Colac	Australian Capital Territory	Batemans Bay
Geraldton			Port Pirie	Greater Adelaide	Bairnsdale	Bowral - Mittagong
Greater Perth			Whyalla	Horsham	Bathurst	Bundaberg
Karratha				Mildura - Buronga	Bendigo	Byron Bay
Port Hedland				Mount Gambier	Burnie - Somerset	Camden Haven
				Mount Isa	Cairns	Coffs Harbour
				Murray Bridge	Castlemaine	Forster - Tuncurry
				Portland	Devonport	Gladstone
				Swan Hill	Dubbo	Gold Coast - Tweed Heads
				Victor Harbor - Goolwa	Echuca - Moama	Grafton
				Warrnambool	Emerald	Greater Brisbane
					Geelong	Greater Sydney
					Goulburn	Gympie
					Greater Hobart	Hervey Bay
					Greater Melbourne	Kempsey
					Griffith	Kingaroy
					Launceston	Lismore
					Mackay	Lithgow
					Moe - Newborough	Maryborough
					Mudgee	Medowie
					Orange	Morisset - Cooranbong
					Sale	Muswellbrook
					Shepparton - Mooroopna	Nelson Bay
					Townsville	Newcastle - Maitland
					Traralgon - Morwell	Nowra - Bomaderry
					Ulverstone	Port Macquarie
					Wagga Wagga	Rockhampton
					Wangaratta	Singleton
					Warragul - Drouin	St Georges Basin - Sanctuary Point
						Sunshine Coast
						Tamworth
						Taree
						Toowoomba
						Ulladulla
						Warwick
						Wollongong
						Yeppoon

Older Coordinate Reference Systems

These are generally no longer used, but are included for completeness.

EPSG	Name	Extent	Notes
3112	GDA94 / Geoscience Australia Lambert	Australia	Older projected CRS that's useful for Australia-wide distance calculations
3111	GDA94 / Vicgrid	Victoria	Older projected CRS for Victoria
3308	GDA94 / NSW Lambert	NSW	Older projected CRS for NSW
3107	GDA94 / SA Lambert	SA	Older projected CRS for SA