Skip to main content

Last reviewed: 26 August 2026

Things GIS cannot represent well

GIS is very good at answering questions such as where, how far, how many, what overlaps and what changed.

It is much less reliable when the thing being represented is relational, contested, seasonal, sacred, embodied, remembered differently or not meant to be separated from the people who hold the knowledge.

Recognising those limits is not anti-technology. It is good GIS practice.

Whakapapa​

A network can draw connections among people, places and events. It cannot carry the full meaning of whakapapa.

The diagram may show a relationship while missing:

  • the responsibilities that come with it
  • differing whakapapa lines
  • the context in which the relationship matters
  • what should remain within whānau
  • the kōrero that gives the relationship meaning

Use GIS as an index or supporting view rather than assuming the diagram has become the whakapapa.

See Whakapapa and mapping.

Overlapping relationships​

GIS software likes clean polygons. Māori relationships with place may overlap.

A single line can hide:

  • shared use
  • ahi kā
  • different responsibilities for different resources
  • relationships through several hapū
  • seasonal use
  • historical change
  • different boundaries for different kaupapa

Do not force one geometry to answer every question.

Sometimes several overlapping layers are more useful. Sometimes the answer is a narrative, network or deliberately indicative area.

Kōrero tuku iho​

A point labelled with three sentences can be useful for orientation. It may also strip a kōrero from sequence, voice, place, audience and context.

If the story matters more than the coordinate, useful options include:

  • keeping the full kōrero in an archive or oral-history system
  • using the map only to provide context suited to the audience
  • linking several places to one narrative
  • using audio or video for particular audiences
  • leaving the location outside the map

Seasonal relationships​

A permanent polygon is a poor representation of something that changes through the year.

Examples include:

  • mahinga kai
  • species movement
  • customary harvest
  • access conditions
  • water levels
  • seasonal occupation
  • maramataka-related practice

If time matters, include time. If the practice cannot be represented usefully or meaningfully as time-enabled GIS, keep that detail outside the map.

Differing accounts​

GIS databases often enforce one value per field.

History and memory do not always cooperate.

Two people may remember a location differently. Two sources may use different names. Several versions may all matter.

Do not average disagreement into false certainty.

Useful approaches include:

  • multiple source records linked to one place
  • confidence or status fields
  • narrative notes explaining the difference
  • separate geometries rather than one invented compromise
  • retaining the original source wording in a controlled archive

Wāhi whose significance is not spatially bounded​

Some places cannot sensibly be represented by drawing a polygon around the important bit.

The significance may lie in:

  • a view between places
  • a journey
  • an association across a whole landscape
  • sound, water flow or visibility
  • a relationship among several features
  • an event that cannot be reduced to one point

The geometry should follow the kaupapa, not the convenience of the GIS tool.

Knowledge that stays outside GIS​

Not everything becomes stronger or more useful when digitised.

A project may keep some information:

  • unrecorded in the GIS
  • recorded only as a reference that more kōrero exists
  • in another information system
  • available only in particular settings
  • outside consultant or cloud workflows

A blank space in a database can be deliberate. It does not automatically mean a data-quality problem.

Relationships that do not fit ownership​

Conventional GIS often reaches quickly for fields such as owner, boundary, asset, stakeholder and site.

Those terms may be useful for particular administrative tasks. They may also distort relationships involving:

  • kaitiakitanga
  • mana whenua
  • customary use
  • shared responsibility
  • tūrangawaewae
  • whakapapa
  • collective rights and obligations

Use the field names required for the work, but do not let an administrative schema redefine the relationship.

Uncertainty​

A sharp line can make weak evidence look precise.

Use uncertainty explicitly through:

  • indicative areas
  • confidence bands
  • generalised geometry
  • source-specific geometries
  • date ranges
  • notes explaining what is unknown
  • maps that show several plausible interpretations

If a map reader can see more certainty than the source supports, the cartography is misleading.

Silence and absence​

No feature on the map does not necessarily mean nothing was there.

Absence may mean:

  • the source was never collected
  • the information sits elsewhere
  • a name was lost from later maps
  • the surveyor did not record it
  • a database has incomplete coverage
  • the people with knowledge were not part of that source
  • the project deliberately chose not to digitise it

Do not turn no data into no relationship.

A practical test​

Before designing a new layer, ask:

  1. What will the geometry make appear certain?
  2. What relationships will disappear when the information becomes a feature?
  3. What will a future user assume if they never hear the original kōrero?
  4. Is the map the best representation, or merely the easiest one for the project team?
  5. Could the GIS store only a reference and leave the deeper knowledge elsewhere?

The best Māori GIS database is not necessarily the one that contains the most Māori knowledge. It is the one that holds what the kaupapa genuinely needs while keeping relationships, context and provenance visible.

Last updated on