Campaigns and party research
For party research desks, campaign directors and community-independent campaigns — seats as data, results down to the booth, joined to demographics on the ABS spine across changing boundaries.
Politics
Electorate analysis
Seats as data. Federal and state results down to the booth, joined to census demographics on the ABS spine, across boundary sets that keep changing.
Every campaign, party office and peak body in the country keeps the same spreadsheet. Seats down one side, margins and demographics across the top, assembled by hand from tally room downloads and census tables. It takes weeks, one person owns it, and a redistribution turns it into fiction overnight.
South Australia, Tasmania and the Australian Capital Territory are being redistributed federally right now. NSW state boundaries are redrawn before the 2027 election. Whoever maintains that spreadsheet is about to rebuild it.
The join that makes it work
Election results are published by polling place and division. Demographics are published by mesh block and statistical area. Neither knows about the other, and the boundaries move independently — divisions on the redistribution cycle, statistical areas on the census cycle.
The bridge is the small-area layer. Booth results plus roll mark-off records let votes be apportioned to statistical areas, and statistical areas are the currency everything else in the lake already speaks: census counts, socioeconomic indices, population estimates, industry of employment, and every hazard and land layer used elsewhere on this site. Once votes sit on that spine, a seat profile is a query rather than a project.
Rebuild the notional margins for the proposed South Australian divisions from the 2025 results, and list which suburbs moved between seats.
The output is a table of divisions under the proposed boundaries with the notional two-party-preferred result, the change against the current boundaries, and a companion list of the areas that moved with their population and demographic profile. The method — how votes were apportioned, which boundary vintages were used, where the apportionment is weakest — is written out alongside it, because a notional margin without its method is just an assertion.
Seat profiles and shift
The second question everyone asks is what changed. Census to census, a division's age structure, income, industry, housing tenure and language profile move, and the direction of that movement is usually a better predictor than the level.
For these twelve marginal seats, show the demographic shift from 2016 to 2021 and how much of each division sits in a flood planning area or bushfire prone land.
That second half is the unusual one. Because hazard and climate layers sit on the same spine, exposure can be cut by electorate as easily as by council area — dwellings in a flood planning area per division, projected hot days by 2050, canopy cover and heat, coastline change. Campaigns that run on climate and community independents in particular ask for that cut and can rarely get it.
Coverage
Federal results from 2004 onward, boundary sets including proposals under consultation, and NSW state and local government results are the first build. Derived from those are three marts: a division profile that time-aligns demographics across boundary sets, booth catchments, and a redistribution difference showing which voters moved and what they look like.
The electoral layers are landing now. Everything they join to — the ABS spine, the census, the hazard and climate stack — is already there.
Frequently asked
How do results get to suburb or SA1 level?
Booth results are published by polling place, and the electoral roll mark-off records which small area each voter who attended a booth came from. That link lets votes be apportioned to statistical areas, which is what makes a join to census demographics possible at all. It is an estimate with a stated method, not a count of how any small area voted.
Can it handle a redistribution?
That is the case it is built for. Results counted on old boundaries are re-apportioned to the new ones through the small-area layer, so a notional margin comes with the working shown rather than as a number you have to take on trust.
Is this available today?
The ABS spine, census, socioeconomic indices and population series are live now, and the hazard and land layers that a climate-exposure-by-seat analysis needs are live too. The electoral layers — results, boundary sets, roll-based small-area votes — are being loaded now. Talk to us about timing for a specific state or election.
What about the licence on electoral data?
Australian Electoral Commission material carries its own terms rather than a Creative Commons licence. Derived products and redistribution are permitted with attribution and the required notice, and that notice travels into the dataset documentation and the footer of anything the agent produces from it.
Related
See it run on your portfolio
Zenancy is in private preview with Group 2 reporters and their advisers.
Request access