Skip to main content

Publishing maps with sensitive locations

Some spatial information works differently in a public map than it does in a working dataset. Exact locations can also travel beyond their original audience through screenshots, forwarded PDFs, shared drives and copied layers.

This page describes practical ways to make different versions of a map for different purposes without assuming that one publication rule fits every kaupapa.

Deeper guidance

For more on sensitive places, inference risk, metadata, copies and graduated access, see Sensitive places, Spatial masking, Inference risk and Metadata can give you away.

Sensitivity depends on context​

Sensitivity can arise from sacred places, physical risk, privacy, ecological pressure, commercial value, whakapapa, or simply the fact that information was shared for a particular purpose and audience.

Examples can include:

  • wāhi tapu and wāhi tūpuna
  • urupā and burial areas
  • mahinga kai, rongoā and gathering places
  • taonga species nesting, breeding or habitat sites
  • culturally significant landscapes under pressure
  • access routes, gates, tracks or vulnerabilities on Māori land
  • private whānau locations connected to kōrero
  • information shared for a particular hui, project or working relationship

Match the map to the purpose​

A useful publication question is not simply whether a layer is sensitive. It is what the map needs to communicate to this audience.

Three questions usually get to the point:

  • Who is this version for?
  • What does it need to show?
  • How much spatial detail actually helps?

A public education map, a planning map and an internal working layer can legitimately show different versions of the same underlying information.

Different sharing patterns​

Common patterns include:

  • display during a hui without circulating the working data
  • access within a project group
  • a version prepared for a particular partner or process
  • a public map designed with copying and reuse in mind

The technical platform can then be configured to match the way the particular version is intended to be used.

Ways to reduce unnecessary precision​

Areas instead of points​

A broad polygon can communicate a landscape or area of interest without making one coordinate the focus.

Buffers and distance bands​

A distance band can show that something matters to a planning or operational decision without exposing the underlying point.

Grids and zones​

Points can be summarised into grid cells where distribution matters more than individual locations.

Density surfaces and heatmaps​

Aggregated surfaces can show patterns without displaying every feature.

Counts without geometry​

Counts by catchment, management unit, land block or another area can sometimes do the whole job.

Category labels instead of names​

A shared version can use broader categories where individual site names add little to the purpose.

Separate working and shared versions​

One dataset can retain the best available working geometry and source context while another version is prepared for a wider audience. This is often clearer than permanently altering the working record.

Attributes can reveal location too​

Generalised geometry may still be identifiable through:

  • site names or local nicknames
  • detailed descriptions
  • photographs with recognisable backgrounds
  • attachments with embedded coordinates
  • IDs that join back to another dataset
  • map extent, title or surrounding layers

The useful check is whether those details are helping the audience or simply carrying extra information into the output.

Historic imagery and overlays​

Historic imagery can reveal places indirectly when combined with current basemaps or other layers. A narrative, cropped image, generalised overlay or non-spatial reference may sometimes communicate the same point more clearly.

A simple design test​

Before producing a shared version, ask:

  1. What does this audience need to understand?
  2. Which details make that clearer?
  3. Which details are only present because they happen to exist in the working data?

That keeps the publication decision tied to the kaupapa rather than to a generic rule about what can or cannot appear on a map.

Useful checks for a shared version​

  • purpose and audience are clear
  • spatial detail matches the job the map is doing
  • attributes included are useful to that audience
  • source and date are visible where they matter
  • geometry and labels do not imply more certainty than the source supports
  • photos, attachments, metadata and map extent have been considered as part of the output
  • the shared version is distinguishable from the main working record
  • enough decision context is kept that somebody can understand later why this version was made

What to write on the map​

Plain statements about what the map represents are often more useful than warning language.

Examples:

  • Locations are shown at a generalised level for this map.
  • This map shows landscape context rather than exact site positions.
  • Prepared for the August 2026 planning hui.
  • Source and spatial detail vary between records. See the accompanying notes.

Last reviewed: 25 August 2026

Last updated on