Skip to content

Methodology

How we count.

We say “90+ government datasets”. Here’s the rule and the arithmetic, so you can argue with it.

  1. 112wired into live use
  2. − 16US, UK or NZ only
  3. = 96readable in Australia
  4. 90+what we say

Rounded down on purpose. Under-claiming is the only safe way to be wrong. Dataset figure as at August 2026; census status: provisional.

One dataset, per publisher, per check.

Counts once

NSW ePlanning’s Principal Planning Layers: one, not six.

Counts separately

VIC Heritage Register and Heritage Inventory: two instruments.

NSW coastal SEPP 2.7 · 2.10 · 2.11: different consequences.

Only if a report reads it

Data stored with nothing reading it isn’t a capability. It doesn’t count.

Three numbers. Never blended.

Per report

checks, stated exactly

Every report states exactly how many checks ran at that address, and which ones didn't. The exact checks run depend on the address, state and available coverage.

90+

government datasets

The published sources those checks read. One check often reads several; some datasets serve several checks.

How we test ourselves

The platform is tested as carefully as the property.

These test KRAYN's own software and the reports it has produced.

285

automated tests run on our deploy path. Any one of them can stop it. As at 2026-10-18.

What they will not let a report say:

REFUSED
"Largely clear — a few things to check"
on a report whose own checks found a major issue.
KRAYN
The verdict is raised to meet the worst finding.

“up to 30 site checks on one address in Queensland, across 90+ datasets” reads like one claim. It’s two.

What the count leaves out.

  • Data we hold but don’t read. Excluded by name.
  • Non-government sources. OpenStreetMap, SoilGrids, commercial weather.
  • Overseas layers. They can’t fire on an Australian address.
  • Layers not yet built. Nothing until a report can read them.

The dataset register - the table of when each dataset was last refreshed - is temporarily unavailable.

It was showing an estimated next-refresh date for every dataset, calculated three months after the last one. Those dates were not published by the data owners, and some of them contradicted how often we actually re-read the data. We've taken the register down rather than leave our estimates standing as if they were commitments.

It will come back with the dates we can source, and nothing where we can't.

The real receipt is the citation on every finding.