About · Methodology · Assumptions
Where the data comes from.
Every chart, map, and ratio on this site is stitched together from roughly a dozen external data families. This page inventories each one—by agency, product, vintage, geographic coverage, and what the model does with it—so you can trace any number back to its source.
Snapshot version:
How the raw data is organised
Two fetch families, one SHA-pinned manifest
Raw spatial inputs—about 24 GB—live outside git in each state’s
analysis/<state>/data/raw/ tree and are pulled from the
project’s Cloudflare R2 bucket via
scripts/fetch_raw_data.py, gated by a SHA-256 manifest (scripts/raw_data_manifest.json) so an ingest can’t silently swap an input. A separate family of
fetch_*.py scripts pulls the always-online ABS, Valuer-General,
and GTFS data directly from public CDNs on demand—census DataPacks, SEIFA,
building approvals, regional population, sale prices, and the transit feeds.
The canonical human-readable inventory is
docs/raw-data-inventory.md, filed by Ethan as #92 and
committed next to the code that reads each input. It carries a status
legend—Confirmed URL, Confirmed source, Hand-curated, Derived, Needs
citation—so each source’s provenance confidence is explicit rather
than implied.
Source: docs/raw-data-inventory.md (#92), scripts/fetch_raw_data.py, scripts/raw_data_manifest.json. Methodology →
Read this before pairing numbers
The vintages are deliberately mismatched
The four temporal layers don’t line up—by design, not oversight:
| Layer | Vintage |
|---|---|
| Zoning / restriction | Current (2026) planning snapshot |
| Census income / rent / mortgage / demographics | 2021 Census |
| Building approvals | Annual flow (latest single FY 2024-25; 5-FY history available) |
| Regional population (ERP) | Per-FY current (demographics use 30 June 2025; per-ring population uses an older ~2023 ERP) |
| Sale prices | Mix of Census-2021 imputation + Valuer-General annual files |
The prices README is emphatic: do not pair these as if one moves with
another. A high-price, low-restriction suburb is not proof that recent
reform cut prices—the prices are roughly four years older than the
restriction snapshot. Every demographics row bakes in provenance
columns (population_vintage, dwellings_vintage, *_source) so the consumer can see exactly which
vintage each number is.
Source: outputs/combined/prices/README.md, outputs/combined/demographics/README.md.
Source family 1 of 6
ABS Census 2021—DataPacks, mesh blocks, SEIFA
General Community Profile (GCP) DataPacks supply the demographic and affordability columns; Mesh Block Counts give the finest population and dwelling denominators; SEIFA supplies the socioeconomic indices. All CC-BY 4.0.
| Product / table | Geographic levels | What the model uses it for |
|---|---|---|
| GCP G01—selected person characteristics (age, country of birth, language) | SAL · SA2 · LGA | Median-age estimate, over-65 / under-18 %, Australian-born %, English-only-at-home % for demographic × restriction correlations |
| GCP G02—selected medians and averages | SAL · SA2 · LGA | Median weekly household income, median weekly rent, median monthly mortgage—affordability ratios + price/income regressions |
| GCP G09—country of birth detail | SAL · SA2 · LGA | Top-7 country-of-birth shares (England, China, India, NZ, Italy, Vietnam, Philippines) for migrant-composition cross-cuts |
| GCP G14—religious affiliation | SAL · SA2 · LGA | no-religion % and Christian % demographic columns |
| GCP G15—educational institution attending | SAL · SA2 · LGA | University-attending % (student-precinct signal) |
| GCP G34—motor vehicles per dwelling | SAL · SA2 · LGA | Zero-car-household % + mean vehicles per dwelling |
| GCP G36—dwelling structure | SAL · SA2 · LGA | Separate-house / medium-density / apartment % |
| GCP G37—tenure type × dwelling structure | SAL · SA2 · LGA | Owner-outright / owner-with-mortgage / renter shares |
| GCP G49—non-school qualification / attainment | SAL · SA2 · LGA | Bachelor-or-higher % |
| GCP G60—occupation × age | SAL · SA2 · LGA | Manager-or-professional % |
| Mesh Block Counts 2021—person + dwelling per MB | Mesh Block (MB_2021) | Per-ring population + dwelling counts; existing-dwelling-stock denominator for approvals-per-1,000-dwellings; parcel-level dwelling distribution |
| SEIFA 2021 (2033.0)—IRSAD, IRSD, IER, IEO indices + deciles | SA2 · LGA · SAL | Socioeconomic advantage/disadvantage scores joined to restriction data |
G-tables are fetched by scripts/fetch_abs_census.py from the
ABS DataPacks CDN (short-header AUS-wide ZIPs). The fetcher also extracts
G04/G07/G10/G11/G17/G33/G38/G40/G46/G50/G56/G62 from the same ZIPs, but
several of those were later dropped as “subsumed” or “too granular for a
headline”. SEIFA is fetched by
scripts/fetch_abs_seifa.py; mesh-block counts are read
from the gitignored “Mesh Block Counts, 2021.xlsx”.
Source: ABS Census 2021 GCP DataPacks (CC-BY 4.0), ABS 2033.0 SEIFA, ABS 2021 Mesh Block Counts; scripts/fetch_abs_census.py, scripts/fetch_abs_seifa.py. Methodology →
Source family 2 of 6
Zoning & planning-scheme layers, per jurisdiction
There is no national zoning dataset—each state and territory publishes its own. Queensland publishes none at the state level, so Brisbane is a per-council merge of five schemes. Zone-code→category and height/min-lot lookups are hand-curated CSVs per state, sitting alongside the spatial layers.
| State / territory | Source agency + product | Format | Access | What it provides |
|---|---|---|---|---|
| NSW (Greater Sydney) | NSW DPHI—Standard Instrument LEP Land Zoning (EPI Primary Planning Layers, MapServer/2) | Shapefile (ArcGIS REST) | Confirmed URL | Zone polygons; plus LEP Height of Building (HOB, MapServer/5, ~41,000 polygons, MAX_B_H metres → storeys) |
| VIC (Melbourne) | DTP VicPlan—PLAN_ZONE statewide planning-scheme zones (planning_scheme_zones/MapServer/0) | GeoPackage (ArcGIS REST) | Confirmed URL | Statewide zone polygons; ZONE_CODE (e.g. NRZ1) + ZONE_CODE_GROUP prefix; per-LGA NRZ/GRZ schedule heights hand-coded (#383) |
| QLD (Greater Brisbane) | Per-council—Brisbane City Plan 2014; Ipswich (council-supplied); Logan Planning Scheme v9.2; Moreton Bay; Redland City Plan | Shapefile (ArcGIS Hub / REST / direct council) | Mixed—confirmed URLs + 2 council-supplied | No statewide scheme; 5 sub-councils processed separately then merged; each contributes zones + heights |
| SA (Adelaide) | data.sa.gov.au—Planning & Design Code Zones (+ SubZones + Technical & Numeric Variations) + Land Use Generalised 2025 | Shapefile (GDA2020) | Confirmed URL | P&D Code zones; variations give local height/lot overrides; land-use ∩ zones is the base residential layer |
| WA (Greater Perth) | Landgate SLIP—Local Planning Scheme R-Codes (DPLH-070, MapServer/111) + Perth CPS2 Maximum Building Height | Shapefile (SLIP REST) | Confirmed URL | R-Code density-designator zoning; CPS2 height limits for City of Perth |
| TAS (Hobart) | LISTdata—Tasmanian Planning Scheme Zones (PlanningOnline MapServer/13) + Interim Planning Zones (/4) + Code Overlays (/14, /3) | Shapefile (ArcGIS REST) | Confirmed URL | Statewide TPS zones (+ legacy interim scheme); code overlays carry heritage |
| NT (Darwin) | NT DLPE—NT Town Planning Zones + NT Land Use Mapping Project 2016-current (LUMP) | Shapefile | Zones supplied direct from DLPE (no public URL); LUMP via data.nt.gov.au | Town-planning zones; heights/categories hand-coded from NT Planning Scheme 2020 |
| ACT (Canberra) | ACTMAPi / data.act.gov.au—Territory Plan Land Use Zones + ACT Blocks (parcels) | Shapefile + GeoPackage | Confirmed URL | Zoning polygons + block geometries; zone→category + lot/coverage rules hand-curated from Territory Plan |
As of #377, the low-density / rural / established-zone classification
is assigned at source in each R/cities/<city>.R
adapter and baked into parcels.gpkg, rather than patched
in Python downstream. A legacy per-state notebook pipeline that
carried narrower definitions (#117/#121) has been retired (#406), so
the topline low-density-share.csv and the suburb (SAL), SA2
and LGA numbers now share one source-level definition.
Source: Per-state planning-scheme layers (see table); outputs/combined/zoning/README.md, R/cities/<city>.R adapters (#377). Methodology →
Source family 3 of 6
Cadastral parcels—the denominator layer
Parcel-area denominators are a deliberate choice: zoning polygons clip
through roads and easements, so the headline ratios compute their
denominators from cadastral lot_area_m2 per parcel, not from
raw zone-polygon area.
| State | Source / product | Format | Vintage | Notes |
|---|---|---|---|---|
| NSW | NSW Land Parcel and Property Theme (Spatial Services) | GeoPackage (~1.6M parcels) | Current cadastral cut | Land-parcel level (not strata)—corrects the legacy strata multi-count in ring 1; fetched by fetch_nsw_parcels.py |
| VIC | VicMap Cadastre | GeoPackage (~3M lots) | Current cadastral cut | PS-prefix strata-deduped at load; fetched by fetch_vic_parcels.py |
| QLD | Queensland Digital Cadastral DB—Land Parcel Property Framework (Tenure) | Shapefile (ArcGIS REST) | Current | Statewide parcels joined to all 5 sub-councils |
| WA | Landgate cadastre (SLIP Historical Cadastre WFS) | GeoPackage | 2024 vintage | “2024 cadastral.gpkg” |
| TAS | LIST Cadastral Parcels (LISTdata) | GeoPackage | Current | LIST metadata record |
| ACT | ACT Blocks (data.act.gov.au) | GeoPackage | Current | Block geometries used as parcel proxy; the per-distance-ring lot breakdowns (lot count + median lot area by ring) aren’t computed for Canberra, though its overall + per-suburb lot sizes are. |
| SA / NT / ACT | (leaderboard-rollup caveat) | — | — | SA, NT + ACT still report parcel_count = 0 in the suburb- and LGA-level leaderboard rollups (#119/#120), so those tables fall back to an area floor. Lot sizes + min-lot stats ship for all three; the NT (Darwin) now also runs the full per-parcel lot metrics off its land-titles cadastre. |
Source: Per-state cadastral parcel datasets (see table); scripts/fetch_nsw_parcels.py, scripts/fetch_vic_parcels.py. Methodology →
These parcels also yield the derived lot-size dataset (outputs/combined/lot-size/)—the median and mean area of each city’s existing residential lots, at
city, ring, zone, LGA, suburb and SA2 level. That’s the actual typical
lot (Melbourne ~506 m² up to Darwin ~817 m², rural-zoned parcels excluded),
a separate measure from the UDIA greenfield average (the median new-estate
lot).
Source family 4 of 6
ABS statistical boundaries (ASGS)—shared by every city
Every rollup, clip, and join references the Australian Statistical Geography Standard. All editions sit in the 2021–2025 range; outline GeoJSONs are emitted in EPSG:4326 and simplified in EPSG:3577 (Albers).
| Boundary | Product | Agency / edition | Used for |
|---|---|---|---|
| LGA | LGA_2025_AUST_GDA2020 | ABS—Local Government Areas, 2025 (ASGS) | Council-level leaderboards + detail-page outlines |
| SA2 | SA2_2021_AUST_GDA2020 | ABS—Statistical Area Level 2, 2021 (ASGS Ed.3) | SA2 rollups, approvals, demographics, price regressions |
| SAL (suburb/locality) | SAL_2021_AUST_GDA2020 | ABS—Suburbs and Localities, 2021 | Canonical suburb-level geography for leaderboards + sale-price joins |
| GCCSA | gccsa.gpkg → per-city clips | ABS—Greater Capital City Statistical Area, 2021 | Per-city metro footprint clip; headline scope |
| SUA | Significant Urban Area | ABS (ASGS) | Preferred city-level approvals grain (#349)—excludes peri-urban LGAs that pad GCCSA |
| Mesh Block | MB_2021_AUST_GDA2020 | ABS—Mesh Blocks, 2021 | Population + dwelling counts per ring; finest geometry |
| LGA correspondence | LGA_2021 → LGA_2025 correspondence CSV | ABS (ASGS Ed.3) | Cross-vintage boundary mapping |
Source: ABS ASGS boundary products (LGA 2025, SA2/SAL/GCCSA/Mesh Block/SUA 2021); outputs/combined/boundaries/README.md. Methodology →
Source family 5 of 6
GTFS transit feeds—the rapid-stop buffer
“Rapid” means the frequent-network classification: a weekday 7am–7pm mean headway of 15 minutes or less (matching PTV Frequent Network, TransLink BUZ, Adelaide Go Zone, and TfNSW B-Line). A representative term-time Tuesday is auto-picked from each feed’s calendar. The walkable buffer is a straight-line 800m (roughly a 10-minute walk), not network-walk distance.
| City | Agency / feed | Auto-download? | Vintage / note |
|---|---|---|---|
| Adelaide | Adelaide Metro GTFS (gtfs.adelaidemetro.com.au) | Yes (direct URL) | Latest static feed |
| Brisbane | TransLink SEQ_GTFS (gtfsrt.api.translink.com.au) | Pinned | Pipeline pins the April 2025 pre-strike feed (#95) so the near-PT denominator is normal-service rail, not the reduced strike timetable |
| Canberra | Transport Canberra GTFS (transport.act.gov.au) | Yes (direct URL) | MyWay+ feed; absent from the headline PT CSV at last run (coverage gap) |
| Melbourne | PTV GTFS (data.ptv.vic.gov.au) | Yes (direct URL) | Latest static feed |
| Sydney | TfNSW Timetables Complete GTFS (opendata.transport.nsw.gov.au) | No—sign-up gated | ZIP dropped in manually |
| Perth | TransPerth Spatial Data Access | No—dynamic link | ZIP dropped in manually |
| Hobart | Metro Tasmania GTFS (metrotas.com.au/community/gtfs) | No—session-protected | ZIP dropped in manually |
| Darwin | (no public GTFS feed) | n/a | Falls back to an OSM mode filter (rail/tram/ferry) in pt_proximity.py |
⚠ Brisbane feed is intentionally not the live one
Issue #95 is closed, but it remains a manual pin: the pipeline deliberately uses the April 2025 pre-strike TransLink feed so Brisbane’s near-PT denominator reflects normal-service rail rather than the reduced strike timetable. This does not auto-refresh—if a later feed is adopted, the pin must be updated by hand.
Source: Per-city GTFS static feeds (see table); scripts/gtfs_rapid_stops.py, outputs/combined/pt/README.md. Methodology →
Source family 6 of 6
Approvals, population, sale prices & heritage overlays
Building approvals are the “did the market actually build it?” signal; regional population is the denominator for per-1,000-population rates; sale prices feed the affordability regressions; heritage overlays are one of the three highly-restricted controls. Heritage gradings (significant / contributory / representative) are flattened to “Local Heritage” for cross-jurisdiction consistency, and Aboriginal-cultural-heritage items are excluded.
| Dataset | Agency / product | Vintage / coverage | Used for |
|---|---|---|---|
| Building approvals | ABS 8731.0 Building Approvals, Australia—small-area data cubes (CC-BY 4.0) | FY 2020-21 → 2024-25 (5 FYs; small-area cubes only reach back to Mar 2022) | Dwelling-approval counts (new houses + new other residential) per SA2/LGA/SUA/GCCSA × FY; the “did the market actually build it?” supply-side signal |
| Regional population (ERP) | ABS 3218.0 Regional Population—cubes 32180DS0001 (SA2), DS0002 (LGA), DS0005 (GCCSA) | Per-FY ERP, releases 2020-21 → 2024-25; demographics rollup uses ERP at 30 June 2025 | Per-FY denominator for approvals-per-1,000-population; current-FY population in demographics outputs |
| VIC sale prices | Valuer-General Victoria—Median House by Suburb + Units by Suburb (VPSR time series), CC-BY 4.0 | 2013–2023/2024 annual | Melbourne median sale price + 10y CAGR (houses + apartments) |
| SA sale prices | SA Valuer-General—Metro Median House Sales, CC-BY 4.0 | Quarterly | Adelaide median house sale price (no open apartment data) |
| NSW sale prices | NSW Valuer-General—Property Sales Information (PSI) bulk, CC-BY-ND 4.0 | Annual files; 2014 + 2025 for the 11y CAGR | Sydney per-suburb median (houses + apartments) + CAGR—only the derived per-suburb statistic is published (non-substantial extraction under CC 4.0 §2(a)(4)) |
| NSW state heritage | NSW State Heritage Register curtilage (ePlanning Planning Portal, MapServer/221) | Current | State-heritage classification of Sydney parcels |
| VIC heritage | VicMap HO local heritage overlay (overlays/9) + Victorian Heritage Register (heritage/2) | Current (VHR added under #381) | Local + state heritage flags for Melbourne |
| SA heritage | data.sa.gov.au—SA Heritage Places (polygons + points) | Current | Local (heritagecl L/C) + state (S) heritage |
| WA heritage | Landgate—Heritage Areas (DPLH-089) + WA State Register (DPLH-006 / inHerit) | Current | Local + state heritage for Perth |
| QLD heritage | Queensland Heritage Register (QSpatial) + per-council local overlays | Current | State + local heritage across the 5 Brisbane sub-councils |
| TAS heritage | TAS Planning Scheme Code Overlay (local) + State Heritage Tas / Tasmanian Heritage Register (gpkg) | Current | Local + state heritage for Hobart |
| ACT heritage | ACTGOV Heritage Register (heritage_scoped.geojson, ACTMAPi) | Current | Local heritage only (no separate ACT state register) |
| NT heritage | (zone-coded—HT zone = heritage; no spatial overlay layer) | Current | Local heritage flag for Darwin |
| Road-strip mask (ACT) | ACTGOV Road Reserves (ACTMAPi Open Data) | Current | Masking road strips out of the residential denominator |
Approvals are fetched by scripts/fetch_abs_approvals.py
(eight state Excel cubes, ~1.3 MB); population by
scripts/fetch_abs_population.py; sale prices by
scripts/fetch_vg_sale_prices.py and
scripts/fetch_nsw_vg_sales.py. A parcel is flagged
heritage when its representative interior point (st_point_on_surface) falls inside an overlay polygon, with State winning over
Local—uniform across the seven spatial adapters; Darwin uses
zone-coded heritage.
Source: ABS 8731.0, ABS 3218.0, per-state Valuer-General + heritage layers (see table); outputs/combined/{approvals,population,prices,heritage}/README.md.
Known limitations
Coverage gaps worth surfacing
- Transit feeds: Sydney, Perth and Hobart GTFS feeds are portal-gated and dropped in by hand. Darwin has no public GTFS at all (an OSM mode-filter fallback stands in), and Canberra was missing from the headline PT CSV at the last run.
- Sale prices: open-licence suburb-level price time series exist only for VIC, SA and NSW—the QLD and WA Valuer-Generals don’t publish open data, so Brisbane and Perth have no CAGR. Apartment-vs-house splits exist only for Melbourne and Sydney (SA is houses only).
- Per-parcel metrics: the ACT runs on a block proxy (and SA on houses-only) for some per-parcel metrics; and SA, NT + ACT still report parcel_count = 0 in the suburb/LGA leaderboard rollups (#119/#120), so those fall back to the per-km² metric. The NT (Darwin) now runs on a real land-titles cadastre, so it carries the per-parcel lot metrics.
- Provenance: NT zoning is supplied direct from DLPE with no public URL, and several QLD inputs are hand-coded from planning-scheme PDFs.
Source: outputs/combined/{pt,prices}/README.md, docs/raw-data-inventory.md status legend. Methodology →
The eight capitals
Each city’s source slice
The zoning, heritage and height inputs differ by jurisdiction. This table pairs each capital’s planning-scheme source (from the per-city profiles) with the analysis radius—uniform across all 8 capitals—and the published zoned-capacity baseline that calibrates its capacity figures, where one exists.
| City | Zoning scheme + source agency | Format | Heritage source | Analysis radius |
|---|---|---|---|---|
| Sydney NSW | Standard Instrument LEP—Land Zoning NSW Department of Planning, Housing & Infrastructure (DPHI) | Shapefile (ArcGIS REST) | LEP heritage layer (local) + State Heritage Register (state wins) | 20 km |
| Perth WA | Local Planning Scheme R-Codes Landgate SLIP | Shapefile (SLIP REST) | LGA Heritage Areas (local) + State Register (state wins) | 20 km |
| Brisbane QLD | City Plan 2014 + 4 neighbouring council schemes Per-council (Brisbane CC, Ipswich, Logan, Moreton Bay, Redland) | Shapefile (ArcGIS Hub / REST / council-supplied) | Per-council local + state shapefiles + QLD Heritage Register | 20 km |
| Melbourne VIC | Victoria Planning Provisions—PLAN_ZONE DTP VicPlan | GeoPackage (ArcGIS REST) | VicMap Heritage Overlay (local) + Victorian Heritage Register (state) | 20 km |
| Adelaide SA | Planning & Design Code—Zones data.sa.gov.au | Shapefile (GDA2020) | SA Heritage Places (local = 'L', state = 'S' in one file) | 20 km |
| Hobart TAS | Tasmanian Planning Scheme—Zones LISTdata | Shapefile (ArcGIS REST) | TPS Code Overlay (local) + Tasmanian Heritage Register (state) | 20 km |
| Canberra ACT | Territory Plan—Land Zones ACTMAPi / data.act.gov.au | Shapefile + GeoPackage | ACT heritage register (local); no separate state layer | 20 km |
| Darwin NT | NT Planning Scheme—Town Planning Zones NT DLPE (supplied direct; no public URL) | Shapefile | Zone-coded (TOWNP_ZONE 'HT' → heritage); no NT state register in data | 20 km |
Per-city data quirks
| City | Low-density zone codes | Height source | Quirk / note |
|---|---|---|---|
| Sydney |
| LEP Height of Building (HOB) in metres → storeys (3.1 m/storey) | R5 reclassified to the rural subset (#373). |
| Perth |
| R-Code → storeys lookup; CPS2 height in metres → storeys | Split R-codes (e.g. R20/40) collapse to the highest density. |
| Brisbane |
| Per-council: metres → storeys, or direct storey field | Only multi-source city—5 sub-councils processed separately, then unioned. |
| Melbourne |
| Per-LGA NRZ/GRZ height schedule (metres) → storeys; zone-group fallback | VIC housing zone set carries no farming/rural zones (#370). |
| Adelaide |
| P&D Code 'Maximum Building Height (Levels)' (storeys), or metres → storeys | — |
| Hobart |
| Per-zone building-height storey count (TPS zoning table) | — |
| Canberra |
| Per-zone storey lookup (ACT zoning table; DES sentinel = unlimited) | — |
| Darwin |
| Per-zone storey limit (NT zoning table) | Built from the NT land-titles cadastre (20,588 residential lots, median ~817 m²), so Darwin now runs the per-ring lot metrics, lot-by-lot, frontage and reform-uplift calcs like every other parcel city. About 16% of lots are dense-zone strata that ship a placeholder geometry; they’re kept for area, count and zoning but sit out the geometry-dependent stages, which run on the 17,326 lots with reliable geometry. Heritage is structurally near-zero—the NT data carries only an HT zone, no separate heritage register. |
Source: src/lib/cityProfiles.ts (per-scheme SSOT), cities.json (analysis_radius_km), zonedCapacityBaselines.json (calibration baselines). Methodology →
Capacity calibration—settled
The published zoned-capacity baselines above feed the reform/uplift
capacity calculation, which now runs on the reform-uplift matrix (four
named reforms × the 16-combo grid, net-additional lot-by-lot across
the whole metro). The lot-by-lot capacity method ships alongside the
simple-area method, the yield-methodology audit is closed, and the
in-repo capacity and yield formulas replaced the retired GMAPS model.
Model assumptions—absorption rate alpha, storey factors—are
centralised in config/model-assumptions.json. These
capacity and scenario numbers are the live headline source and safe to
cite. See Assumptions for the model knobs.
Open data
Download the processed files
This page inventories the inputs. The processed outputs—every per-city, per-ring, per-suburb summary the charts read from—are open, citable, and version-stamped. There are 31 per-metric datasets plus the suburb / SA2 / LGA rollups and headline cumulatives.
Browse & download all datasets →
Every JSON is served from
https://zoning.org.au/data/json/ with permissive CORS and a 1-day
edge cache—pull it from the command line, a notebook, or a
fetch call, no auth.
Source: YIMBY-101-Data pipeline outputs/combined; served as CC-BY 4.0 JSON.
Keep reading
Where these inputs become numbers
This page is the “what we ingest”. For the “how we turn it into a figure”, see Methodology (rings, storey conversion, the highly-restricted union). For the model knobs—absorption rate, calibration, scenario assumptions—see Assumptions.
Cite as: Australian Zoning Atlas (). Source-data inventory. zoning.org.au/about/data. Source data from the YIMBY-101-Data pipeline.