# Testing what AI can do with a land listing, open data and human review

> A field note from Landform Research showing how a Trade Me land listing, open LiDAR data and human review became an early architectural design study for a hut on Aotea Great Barrier Island.

- Canonical page: https://www.landformlab.com/research/blog/aotea-method-study.html
- Author: [Matt Strawbridge](https://www.landformlab.com/research/profile.html)
- Published: 24 August 2026
- Updated: 30 August 2026
- Status: Concept design. Not for construction. No consent set or engineering approval.
- Location: Rosalie Bay Road, Aotea Great Barrier Island, New Zealand

## Summary

The study began with a [Trade Me listing](https://www.trademe.co.nz/a/property/rural/auckland/hauraki-gulf-islands/great-barrier-island/listing/6065018741) for 24.3 hectares of bush. Open parcel and LiDAR data were used to compare three possible building areas. Two hut options were then modelled on the selected ridge-top pad, followed by landscape studies, an early drawing package and a cost range.

AI supported repeatable tasks: assembling terrain, running the siting analysis, rebuilding the model, preparing drawings and organising cost information. Matt Strawbridge set the brief, chose the site and design direction, reviewed every output and recorded 28 design and quality decisions.

## What the study produced

- One siting rationale based on slope, sea view, shelter, canopy and walking distance.
- Three candidate building pads at different elevations.
- Two repeatable SketchUp hut models with about 50 named parts each.
- Twelve curated model scenes and exports in glTF, DWG, DXF, PNG and PDF formats.
- Thirteen generated landscape and interior studies, each checked against its source model.
- A nine-sheet Revision E drawing package for specialist review.
- An early cost range built from 31 sourced rates and explicit confidence notes.
- Twenty-eight recorded human review decisions.

## Site evidence and siting analysis

The listing described the property as SE63/NW62 Rosalie Bay Road. The study associated this with Allot NW62 and SE63, Parish of Aotea, title NA573/54. This title information remains subject to independent title confirmation.

Two public one-metre LINZ LiDAR tiles were combined and clipped to the title and its surroundings. The result included matching digital elevation and surface models for the ground and tree canopy. The siting calculation compared:

1. Average slope across a 12 metre building pad.
2. Likely sea-view openness from north-east to south-east.
3. Shelter from the prevailing south-west.
4. Canopy height and likely clearing.
5. Walking distance from the mapped end of Rosalie Bay Road.

The three candidate pads were:

| Candidate | Elevation | Score | Slope | Sea view | Shelter | Canopy | Walk in |
| --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
| A, coastal terrace | 46 m | 0.58 | 14.7 degrees | 0.40 | 0.99 | 10.5 m | 1,953 m |
| B, mid-slope shoulder | 181 m | 0.81 | 11.1 degrees | 1.00 | 0.65 | 7.5 m | 1,821 m |
| C, ridge top, chosen | 243 m | 0.84 | 12.7 degrees | 1.00 | 0.57 | 2.8 m | 1,608 m |

Pad C was selected because it combined the open sea view, the shortest walk and the lowest canopy. It gives up much of the shelter offered by pad A.

## Repeatable model and two design options

The terrain and both hut options were built in SketchUp from repeatable instructions. A local socket connection sent generated Ruby operations to the model. The practical purpose was revision: after each review, the model could be rebuilt and checked again. Drawing views, CAD exports, web geometry and render cameras all came from the same model.

Scheme A is a compact gabled hut drawing on the familiar character of a DOC hut. Scheme B is a taller shutter box whose front weather screen opens to form a canopy.

Scheme B uses one precedent idea from [Hut on Sleds](https://crosson.co.nz/hut-on-sleds-whangapoua/) by Crosson Clarke Carnachan Architects: a weather-closing facade that opens to form an awning. The Aotea study developed a different site, proportion, plan, structure, window arrangement and two-leaf mechanism. Crosson's shutter is one hinged leaf; the Aotea study uses two leaves folding on a head hinge and lifted on twin lines from one common shaft.

## Images

The model established the geometry and camera for each image. Image generation then tested how the options might feel among ponga, flax and coastal bush. The outputs are atmosphere studies, not promises of a finished building. Every frame was checked against its source model for form, openings and major objects.

## Drawing package and specialist review

The Revision E package contains nine A3 sheets, O2-S00 to O2-S08, dated 23 August 2026. The sheets were generated as SVG at 1:5 and include cut outlines, material hatching, hidden work, product notes, dimensions and a scale bar.

The package is explicitly marked **pre-H2 engineering basis, do not fabricate**. Four review checks were closed: the live model was reconciled, polymer was removed from the primary wind restraint, temporary props were deleted and the lift was changed to a synchronised common shaft.

Five checks remain open:

- Geometry arithmetic and the final open projection.
- Final mass and hardware allowance.
- Kinematic path through the dead point and unequal-link offset.
- Operating safety, open-lock reactions and recovery procedure.
- Structural capacity based on confirmed site wind actions and CPEng design.

[Download the Revision E drawing package](https://www.landformlab.com/research/blog/assets/aotea-study/aotea-o2-storm-shutter-rev-e.pdf).

## Early cost range

The estimate uses 31 sourced rates with low and high values and a confidence note. It separates freight, island labour, off-grid plant and consent fees from the building work.

- Scheme A: NZ$183,000 to NZ$401,000, self-framed.
- Scheme B: NZ$224,000 to NZ$484,000, self-framed.
- Self-framing lever: NZ$37,000 to NZ$58,000.
- Lowest-confidence items: access track, barge freight and shutter hardware.

The cost range is directional. The access track, barge rate and specialist shutter assembly require direct quotes.

## Human review

The model and page record 28 review decisions. Examples include rejecting a conventional gable because it felt too much like a house, raising the awning to clear the eyeline, deleting unresolved props, rebuilding walls around real openings, correcting floating vegetation and adding measurable drawing information.

The purpose was not to remove the designer. It was to make repeatable production work fast enough that a person could spend more time comparing, rejecting and refining options.

## Limits and next steps

This is an early concept study prepared to decide whether the land and design direction are worth pursuing. It is not legal, planning, geotechnical, engineering, consent or construction advice. Before progressing, the project requires:

- Independent confirmation of legal title and access.
- A site visit.
- Planning advice.
- Geotechnical investigation.
- Structural engineering by a Chartered Professional Engineer.
- Direct supplier, fabricator, freight and access quotes.

## Related material

- [Visual method study](https://www.landformlab.com/research/blog/aotea-method-study.html)
- [Aotea Field Notes, a 28-page digital volume](https://www.landformlab.com/research/book/aotea.html)
- [One project, three ways to understand it](https://www.landformlab.com/research/blog/sharing-the-work.html)
- [Landform Labs applied work and business enquiries](https://landformlab.com/)
- [Landform Research](https://www.landformlab.com/research/)