APRIQ assembles where housing demand is coming from, how fast supply is actually being delivered, and how rents are responding — population, migration, dwelling approvals, construction and achieved rents, for 1,657 markets. Every figure says what kind of evidence it is, which geography it describes, when it became knowable, and how good the evidence is — with the publisher named wherever its licence asks or naming it helps.
A market is not a rent. Understanding one means knowing how many people are arriving and where from, how much housing is being permitted, how much is actually being finished, and what tenants are paying — on the same page, on their own clocks, each labelled with the geography it was measured for.
Estimated resident population and its three components — natural increase, net internal migration, net overseas migration — from the Australian Bureau of Statistics. Published together, because one of them alone invites a conclusion the other two would change.
Dwelling approvals, commencements, dwellings under construction and completions. An approval is permission to build, not a home: APRIQ shows the whole chain from permission to a place somebody can live in, and never treats one as the other.
Achieved rents from state government bond lodgements — what tenants actually started paying, not what was advertised — for suburbs, postcodes and council areas, with the quartiles behind each median.
A suburb is not a statistical area and a postcode is not a council. Every figure says whether it was measured for the market you are looking at, apportioned into it by an official correspondence, or belongs to a wider area containing it.
Every figure below is a median weekly rent, taken from rental bonds lodged with a state government — what tenants actually started paying, not what was advertised. Five states publish a series we are licensed to republish, and three of them publish a figure for the state as a whole.
Western Australia, the Northern Territory and the ACT have no rent series we can republish yet. What we hold for them is the 2021 Census — one night in August 2021, which cannot show a trend. They are listed on the Australia page.
APRIQ is being built to answer four questions in order, and to be checkable at each one. The first two are live today. The third is under construction behind published validation gates, and the fourth is the reason the archive exists: a forecast made in 2026 can only be graded in 2027 by someone who kept what they said and when they said it.
APRIQ publishes no forecast it has not already committed to grading in public. The scorecard holds the record, ours first, and is empty of claims we have not yet made.
The constraints are the product. Four of them do most of the work; the method page has the rest, the source register and the licences, all generated from the same database as the figures.
Every source file is stored by content hash exactly as retrieved, before anything reads it, and never overwritten. Re-pulling each quarter turns into revision detection — we can show what a figure said before it was revised.
Pages read from a database view that excludes any source whose licence is not confirmed. Data we hold but may not publish is invisible here by construction, not by anyone remembering.
Confidence measures how much evidence a figure rests on. A separate coherence check measures whether it makes sense at all. A median equal to its own third quartile is quarantined even on 200 sales.
Source, licence, retrieval date, file hash, period, method and classification travel with every number — including whether it was observed or calculated.