Sources and method
This site publishes nothing that cannot be traced to a named NDIA table. Where a figure is derived rather than published, it is flagged as an estimate everywhere it appears, and its derivation is set out here.
Sources
Each file is downloaded fresh, content-hashed, and recorded with the hash of the exact bytes the published figures were computed from. The NDIA changes the numeric media ID on every release, so the pipeline resolves IDs by matching link text on the NDIA index pages and fails loudly rather than falling back to a stale file.
| Dataset | Format | Size | Retrieved | SHA-256 |
|---|---|---|---|---|
| SDA Enrolled dwellings and NDIS demand data | xlsx | 1.1 MB | 2026-08-29 | e15d4957b896… |
| SDA participants data | csv | 67 KB | 2026-08-29 | 2ceb31a9f63d… |
| Participants by SA4 data | csv | 109 KB | 2026-08-29 | 9a0f4538fd30… |
| Participants by SA3 data | csv | 361 KB | 2026-08-29 | dfe9c4ddad01… |
| Participants by LGA data | csv | 472 KB | 2026-08-29 | 24001fdea0f3… |
| Participant numbers and plan budgets data | csv | 6.8 MB | 2026-08-29 | 460eb3e7d73a… |
| SIL participants data | csv | 73 KB | 2026-08-29 | 3a5acbc49514… |
| Service District to LGA 2021 data | csv | 23 KB | 2026-08-29 | 30a2b43c5e73… |
| Projected participant by service district data | csv | 9 KB | 2026-08-29 | b3265dc1317a… |
| Specialist Disability Accommodation Report 2026 | 1.6 MB | 2026-08-29 | c6b042e335d9… | |
| NDIS Pricing Arrangements for Specialist Disability Accommodation | 1.2 MB | 2026-08-29 | b53d8df0f282… |
Derived metrics
Anything the NDIA does not publish directly is computed here. Each of these renders on the site with an est marker and its caveat attached.
Enrolled resident places
Supplement P Table P.6
Σ (dwellings in each maximum-residents band × that band's capacity)
The top band is “6+ residents” and is counted as exactly 6, so any dwelling enrolled for more than six residents is undercounted. A dwelling's maximum residents is what it is enrolled for, which may exceed how many people live there.
Implied occupancy
Supplement P Tables P.9 and P.6
participants living in SDA ÷ enrolled resident places
The most contestable number on this site. It divides a participant count by a capacity count taken from a different table. The NDIA states that enrolled capacity “may not accurately reflect current SDA participant occupancy”. It is not a vacancy rate, and it is not an NDIA figure.
Implied unoccupied places
Supplement P Tables P.9 and P.6
enrolled resident places − participants living in SDA
Carries every caveat that implied occupancy carries.
Utilisation
Supplement P Table P.9
participants living in SDA ÷ all participants with an SDA need
Both terms come from the same table, so this is the least contestable of the derived figures.
Share of mix (demand vs supply)
Supplement P Tables P.5 and P.10
each design category ÷ the total across the four categories reported on both sides
Supply is counted in dwellings and demand in participants, so the two are compared only as shares, never as counts. Basic and Multi-Design Category exist only on the supply side; “not recorded” only on the demand side. All three are excluded from the comparison and the excluded share is stated on the chart.
Pipeline
Supplement P Table P.8
unfinished (unenrolled) new build dwellings registered with the NDIA
Not a construction status — the NDIA does not publish one. The Agency gives no guarantee that dwellings listed will be enrolled as SDA.
SDA price limit
NDIS Pricing Arrangements for SDA
base price × location factor
Reproduces the published arithmetic. Legacy stock is excluded because its price table is indexed by resident count and has no location-factor column.
Suppressed values
The NDIA suppresses small counts to protect participant privacy,
publishing them as <11 and occasionally <20 or <30.
- Suppressed cells are carried through the pipeline as a distinct sentinel — never converted to zero, never to blank.
- They render exactly as the NDIA published them.
- They are excluded from sums. Any total that includes one is marked ≥ and described as a lower bound.
- This is why some series here sit slightly below the NDIA’s own national headline. As at 30 June 2026, summing published district figures gives a national SDA in-use count below the 16,644 the Agency reports, because the suppressed districts are missing from the sum. Headline tiles use the NDIA’s unmasked national figure; series summed from districts are labelled as floors.
Geography
The NDIA publishes regional data against 2016 ASGS statistical area codes. The boundary files drawn on the map are the ABS 2021 release, which is the current published edition. Joining one vintage to the other without checking would be a real error, so it was checked.
Every one of the 88 SA4 regions and all 333 matched SA3 regions the NDIA reports carry an identical code in both vintages, so the join is exact for every region shown. The SA3s that do not match are Christmas Island, the Cocos (Keeling) Islands, Jervis Bay, Norfolk Island and Urriarra–Namadgi — none of which have enrolled SDA. Two SA3s appear in the supply workbook but not the participant file (Blue Mountains – South and Illawarra Catchment Reserve), both effectively unpopulated catchment reserves.
Service districts are the NDIA’s own administrative geography and do not nest into ABS areas. They are reconciled to a canonical key so that the SDA participants dataset (“TAS North”) and Supplement P (“TAS – North”) resolve to the same district, and anything that fails to reconcile is reported by the build rather than dropped.
Refresh and changelog
NDIA releases land roughly six to ten weeks after the end of a quarter, on no fixed day. The pipeline polls weekly, hashes every file, and does nothing when nothing has changed. When something has changed it rebuilds and opens a pull request rather than deploying, so a schema change is seen by a person before it reaches the site.
Published periods
- 30 June 2026 14,235 dwellings · 16,644 living in SDA · 6,544 in pipeline
The NDIA replaces its supply tables at the same URL each quarter and publishes no back series for them, so SDA Atlas archives each release as it appears. This list grows by one entry per quarter from the first run.
What the build checks
The publish step fails, and the site keeps last quarter’s numbers, if any of these break:
- Headline figures match values verified independently against the NDIA Quarterly Report.
- Every table’s component columns reconcile to its published total.
- Every reporting date parses, and the newest is not older than what is already published.
- National participant totals move less than 5% quarter on quarter.
- Row counts stay within 20% of the previous release.
- At least 95% of workbook regions join to the participant geography.
- No suppressed value has been turned into a number.
- The pinned pricing document is still the edition expected.
Licence, attribution, disclaimer
All NDIA material used here is published under Creative Commons CC BY 4.0, which permits reuse including commercial use, with attribution.
Contains data © National Disability Insurance Scheme Agency 2013, licensed under CC BY 4.0.
NDIA data disclaimer. This data is provided for the purpose of public insight into the operation of the NDIS. It should not be treated as a forecast. The National Disability Insurance Agency accepts no liability for the representation of this data.
SDA Atlas is independent. It is not affiliated with, endorsed by, or connected to the National Disability Insurance Agency. Errors here are ours, not theirs — every figure can be checked against the linked source above.