If you are researching your first investment property, someone has told you to "check the data". Search ABS census data property investment and you land on the Australian Bureau of Statistics, where the authoritative version lives. The geography behind its numbers covers the whole of Australia, without gaps or overlaps. And before it shows you anything about an area, it asks you to choose a geography, from levels with names like SA1, SA2 and Mesh Block. Most people close the tab at that point. This guide gets you past it: what the Census measures, how its geography works, where to pull the numbers, what the headline figures say, and the traps in reading them. It won't tell you which suburb to buy; it will make the numbers readable before you lean on any of them.
What the Census is, and when it was taken
Every five years the ABS counts every person and house in Australia. The count is a snapshot of a single night: Tuesday 10 August 2021 for the most recent released Census, and Tuesday 11 August 2026 for the one just taken. Everything in it answers one question: where were you, and who was with you, on census night?
Two consequences matter for investment decisions. First, the data ages. The released figures describe 2021, and the data from the 2026 count starts releasing from June 2027, so until then 2021 is the latest Census there is. Second, the geographic boundaries attached to the data are the ones that applied on census night in the year it was collected, which means a suburb's numbers can sit on boundaries that have since been redrawn.
The geography levels, and which one answers your question
ABS geography nests like a set of boxes, and picking the right box is most of the job.
Mesh Blocks are the smallest unit, most containing 30 to 60 dwellings. Statistical Areas Level 1, SA1s, are built from whole Mesh Blocks and are designed to maximise the geographic detail available from Census data; there are 61,845 of them covering the whole country, and they generally hold a population of 200 to 800 people. SA1s aggregate into Statistical Areas Level 2, SA2s, which generally hold between 3,000 and 25,000 people and are drawn to represent a community that interacts socially and economically, which in cities means a suburb. Above that, SA4s reflect labour markets and mostly hold more than 100,000 people. Suburbs and Localities, SALs, are the ABS's approximation of official suburb boundaries, and there are 15,353 of them.
The practical translation: if your question is about a suburb, SA2 or SAL is the level. If you want to separate one pocket of a suburb from another, SA1 is the finest useful cut.
Two derived products sit on these geographies and are worth knowing by name. SEIFA ranks areas by relative socio-economic advantage and disadvantage using Census data, and is published at everything from Local Government Areas down to SA1s, suburbs and postcodes. The Census-based mortgage and rent affordability indicators are published for LGAs and SA2s, with tables for 2011, 2016 and 2021.
This is also the level we work at. PropSpotter's research stage draws on ABS Census data at SA1 level, among more than 30 data sources, and distils the underlying sources into a shortlist of suburbs and a coached decision. The Census is one layer of suburb research rather than the whole job, and our guide to how to research suburbs for investment property walks through the wider process.
Getting the data out: QuickStats and Community Profiles
The ABS publishes Census data through its Search Census data tool, where you select an area and the geography you want it reported at. Two outputs matter. QuickStats are the summary statistics for the area. Community Profiles are the deeper cut: multiple data tables covering demographic, household and dwelling characteristics, downloadable as a spreadsheet. When a variable's meaning is unclear, the 2021 Census dictionary defines the classifications and variables.
The trap at this stage is volume. If you are comparing a handful of suburbs, you do not need every report the suite generates; you need the signals that move the decision. Cutting the data down to the suburb-level numbers that actually drive a purchase, rather than the full firehose, is exactly what our research stage is built to do. If you want that framework spelled out, our guide to the suburb data signals that predict capital growth sets it out step by step.
What the headline numbers say
The 2021 Census counted 10,852,208 private dwellings. Of those, 70 per cent were separate houses, 13 per cent townhouses and 16 per cent apartments. Among occupied dwellings, 31 per cent were owned outright, 35 per cent were owned with a mortgage and 30.6 per cent were rented. And 1,043,776 dwellings were unoccupied on census night, a figure that deserves its own section below.
The Census counts dwellings; for the investor side of the market, the Reserve Bank's dataset on housing investors fills the picture. As at 2022/23 there were 2.3 million individual housing investors, roughly 10 per cent of the working-age population, with investment properties accounting for around 20 per cent of the dwelling stock. Around 70 per cent of those investors owned just one investment property, and among the investors who held several, around 80 per cent held them within a single state or territory. Read together, those numbers say one property is the norm rather than the training wheels: most investors never go past the first, and the ones who do mostly stay close to home.
The unoccupied-dwellings number, read properly
The unoccupied count is the headline figure most often misread, and the ABS says so itself. A property recorded as unoccupied on census night does not necessarily mean it is permanently vacant or available for housing; homes sit empty for many reasons, including holidays, travel, temporary absences or being between occupants. The ABS treats the count with the same care it asks of owners: it asks anyone who owns a holiday home or investment property to report it if the property was vacant on census night. The information is collected only to ensure accurate Census statistics, not shared with anyone, including other government agencies.
So take the 1,043,776 for what it is: a one-night count, not a vacancy survey. Holiday homes, travel and properties between tenants all sit inside it. Anyone quoting it as "one in ten homes sits empty" is reading a snapshot as a stocktake.
Between Censuses: the series that keep moving
Five-yearly data needs companions, and the ABS runs them. Total Value of Dwellings is updated quarterly, estimating the total value, number and mean price of Australia's dwellings. The preliminary estimate for the June quarter of 2026 put the total at $12,688.9 billion, across 11,531,100 dwellings, with the mean price of residential dwellings at $1,100,400 after an $8,200 fall in the quarter. The state spread is wide: New South Wales's mean of $1,304,900 remains the highest in the country, ahead of Queensland at $1,130,600 and Western Australia at $1,123,700, with the Northern Territory lowest at $614,400. Building Approvals is published monthly, covering the number of dwelling units approved and the value of buildings approved.
The division of labour is the point. The Census tells you a suburb's structure, its tenure mix, its housing stock and its households, as at one night every five years. These series tell you where prices and supply sit now. Use both, and know which one you are looking at when a number settles an argument.
How PropSpotter turns this data into a shortlist
PropSpotter was built for people who have never bought an investment property before, and the research stage is where this data work happens. The research brief draws on more than 30 data sources, including ABS Census figures at SA1 level, supply and demand indicators, rental yields, infrastructure pipelines and growth modelling, delivered as a comprehensive PDF brief. The sourcing engine monitors your target suburbs 24/7 and notifies you the moment listings go live, screening each one automatically for bushfire risk, public housing density and owner-occupier ratios. Coaching runs one-on-one with an experienced investor over WhatsApp, covering deal assessment on comparable sales, rental yield projections and risk factors, your negotiation strategy before you make an offer, and a post-purchase debrief after settlement. Against the checklist this guide has given you: the brief works at SA1 level across 30-plus sources, and it filters the data down to the numbers that drive a purchase rather than the full firehose. The whole system is a fixed $4,990 with no commissions and no hidden fees, and it covers all Australian states.
If you want a data-backed read on the suburbs on your own list, start with a free 30-minute strategy call and one-on-one property investment coaching.