Skip to main content
lab-notes.aiPublic-service AI playbooks
Opportunity 04/17EnvironmentEnvironment, land and water

Earth Observation for Public Services

Explore what a mapped land-cover change would need to show before an analyst could act, using survey squares and broad habitat classes.

Inspect starter dataSet up the build partner
Published sources
2
Starter data
20 synthetic records
Domain build partner
build-earth-observation
Last reviewed

From the draft strategy

Opportunity

The draft strategy names AI analysis of satellite data as a potential public-service application, mentioning deforestation, land use, and coastal erosion, without choosing the imagery, the scale, or what anyone would do once a change was mapped.

Where the draft says it

Table 2 — potential public-service applications

Read the draft strategy source (opens in a new tab)

Source register

Research already done

Real published source

We reviewed these published sources so you do not have to start from zero. Each source shows what it covers, how it can be accessed, and why it matters.

2 sources

  1. Department of Agriculture, Environment and Rural Affairs

    Open

    Northern Ireland Countryside Survey

    What it covers
    A repeated field survey of a random sample of 500 by 500 metre squares — 288 of them, about half a percent of Northern Ireland — mapping land cover and habitat by type and comparing each round with the last, from the baseline in the late 1980s to the 2023/24 cycle.
    Why it matters here
    It is the published measure of how land cover changes here, and it is fieldwork rather than imagery, which is the comparison any satellite-based claim in this example would have to face.

    Open the source (opens in a new tab)

  2. Copernicus Land Monitoring Service

    Open

    CORINE Land Cover

    What it covers
    Europe-wide land cover and land cover change inventories built from satellite imagery against a fixed list of classes, free to download and free to use for any purpose.
    Why it matters here
    It is the openly licensed imagery-derived layer this example could genuinely start from, and its class list and smallest mapped area decide which changes it can pick up at all.

    Open the source (opens in a new tab)

Starter data

Starter dataset

Synthetic working data

Use 20 synthetic change rows across ten survey-sized tiles to explore land-cover change without imagery or licensing questions.

AI-assisted research helped identify and interpret the published sources. We then created a small, non-sensitive synthetic dataset shaped by the information those sources expose. It is for exploration—not evidence, training, or operational decisions.

Records
20
Fields
5

Field preview

table view
  • id

    SYN-EO-01 · SYN-EO-02 · SYN-EO-03

  • tile

    Tile 01 · Tile 02 · Tile 03

  • period

    1998-2007 · 2007-2024

  • landCoverClass

    Improved grassland · Broadleaved woodland · Bog

  • changedHectares

    1.8 · 0.6 · 0.4

Inspect all 20 recordsView repository file (opens in a new tab)Download JSON (opens in a new tab)

How it was prepared

AI authored fictional lettered tiles with one broad habitat class and hectares-changed figures for two periods.

Limitations

  • The tiles are numbers and the hectares are invented, so no row here says anything about land anywhere in Northern Ireland.
  • A hectares-changed figure has already thrown the picture away: cloud, shadow, tide, season, and alignment error are what a real pipeline argues about, and none of them survive into a table.
  • One class per tile is a fiction — a real square holds a mosaic — and where one class ends and the next begins is itself a judgement.

Clone and build

Domain build partner

Build with a domain-aware coding agent

Clone the repository, then ask your coding agent to load the skill that ships with it. The skill brings the opportunity, sources, starter data, known unknowns and constraints into the conversation before anything is proposed. It follows the open Agent Skills standard, so it works in Claude Code, Codex and any other agent that reads skills.

01 / Clone

git clone https://github.com/Hypership-Software/lab-notes.ai.git

02 / Invoke

Claude Code
/build-earth-observation
Codex
$build-earth-observation

Conditions, not a checklist

Before you build

These are not footnotes. They are conditions that any responsible prototype would need to address.

  • Wrong classes misdirect action

    A wrong class sends inspection, planning, or an intervention to the wrong field, and the map goes on looking authoritative either way.

  • Detailed imagery can expose sensitive sites

    Fine-grained imagery can expose sensitive habitats and sites, so what gets published needs deciding separately from whether the analysis works.

  • Surface change does not explain cause

    Surface change does not explain cause, ownership, or legality, and a coarse image can miss a small change that matters more than a large one.