Choosing Geography (AU)
(Or where do you want to know about)
Choosing what type of geography you want to get data for can be highly confusing. It's a big subject so I'll only be able to cover a few of the basics in a single page. If you want the full details just reach out and I'll explain what I know and where to find out more.
​
For Vital Signs reporting a combination of Local Government Areas (LGAs) which are your councils, complimented by SA2 and UCL (descriptions below) are normally sufficient.
We cover mainland Australia + Tasmania (NZ is coming)
Right at the top of Australian Geography is our country, for Citizen Data we consider this to be the mainland of Australia and Tasmania, which excludes places like Norfolk Island, if that's a problem we can run the data, but for the time being we're looking just at mainland and Tassie.
​
Typically in your data you'll want numbers from your local area, and also State/Territory and All Australia. So that's what we deliver, typically Australia will be first, the relevant State will be second, and then specific areas, regions, or geographies of interest for your reports and analysis.
​
Different Geographies
The ABS data for the 2021 Census can be delivered using any one of 17 different Geographies. ranging from whole of Country at the highest level where each fact from the Census has 1 line of values, all the way down to Statistical Area 1 (SA1) at the most granular, where each fact has 61844 lines in the dataset. The ABS does a fairly rubbish job of explaining the geographies, so in many instances you have to look at the data, and how it's overlayed onto a map to get an idea of what each one is.
One aspect is important to understand, whichever geography you're going to use, then the areas never overlap. So in our 60K SA1 'tiles' each one is unique and seperate from the others, rather like a very crinkly jigsaw puzzle.
Postcodes (Least Good)
A lot of data is available at the PostCode level, and certainly all the Census's can be delivered using this. However, other than datasets where this is the only option I like to avoid using Postcodes where possible.
The area of a Postcode can and does change frequently, they're oriented toward what is convenient for delivering letters not what would be useful for data and analysis. Nearly all Postcodes are known by multiple names and it can be confusing where the same name is used in different states. There are 2644 individual numbered postcodes, and around 17,000 postcode names.
Local Government Areas - LGAs
An LGA is a council, we all live within an LGA and they are relatively stable boundary wise. For those reasons I prefer using LGAs where possible. They have good name recognition for communities, and they are also one of the main ways of delivering data from sources other than the Census. There are 566 LGA codes, which equates to about 550 councils.
The behemoth is Brisbane which is one single council containing 1.25M people (LGA code 31000), the next biggest is half that size at 625K people. At the other end of the spectrum there are 73 LGA codes with less than 1000 people. So certainly the range in population size and usefulness of LGA data is variable, we can advise what will work best for your report.
​
Typically you'd be using one or more LGAs, local councils who are interested in your (Vital Signs) report, and those geographies would be combined with others.
Urban Centres and Localities - UCLs
These are misnamed as they're not really urban areas more often townships in rural and regional Australia. As such they're very useful for looking in detail at small places, often with populations in the 100-500 range. There are 1836 individual UCLs and although they do change overtime, many have been consistent in area, if not in code number, over the last 4 censuses.
Statistical Areas (SA1 to SA4)
These geographies are created and defined by the ABS themselves, they form a hierarchy with the SA4 at the top with 107 different codes, that equate to the major cities in Australia, and then broad rural and regional parts as well. 83 of the SA4 regions have over 100,000 people, the largest has 860K.
​
A complete subset of SA3 regions fit within a single parent SA4, and in turn a complete subset of SA2 regions fit within an SA2, SA1 within SA2 and the very lowest level used by the ABS is the meshblock with AFAIK are not generally made available to the public as the size of the geography is so small that it could easily identify individuals.
​
​
