How to Ask a Clear Spatial Question
A clear spatial question tells a Geobble Project what outcome you need, where it applies, which Sources should be used, how the relationship should be measured and what form the result should take.
Provide a publication-ready procedure for “How to Ask a Clear Spatial Question”, including prerequisites, expected result, review points, and likely recovery paths.
How to Ask a Clear Spatial Question
A Geobble Project is a conversational workspace, but a good spatial request still needs enough structure to be reviewable. State what you want to know, which Sources should be used, where the question applies, how the relationship should be measured, and what result you want back.
The goal is not to write a perfect prompt, but to avoid making the system invent assumptions that materially change the analysis.
Before you ask
Open or create the Project for this question.
Add the relevant Sources or Maps as context.
Preview the inputs and confirm their geography, dates, fields and geometry types.
Separate unrelated objectives into different requests when they would require different assumptions or outputs.
Projects can support work such as filtering, joins, buffers, intersections, routing, accessibility analysis, comparisons, derived layers and map drafting. The request should make the intended operation understandable enough to review before execution.
Use a five-part question
A useful pattern is:
Using [inputs], calculate or identify [relationship] within [geographic scope], using [method, units or threshold], and return [output]. Treat [important exception] as [rule].
For example:
Using the clinics and districts in this Project, count clinics within each district, keep districts with zero matches, and return a derived polygon Source with the district identifier and clinic count. Do not treat a missing clinic geometry as a zero-count observation.
The five parts are:
Inputs — name the intended Sources or fields rather than saying “the data”.
Question — say what relationship or measure you need.
Geography — identify the study area and boundary version when it matters.
Method — define distance, travel mode, units, threshold, predicate, denominator or other consequential rule.
Output — say whether you need a preview, table, derived Source, map or summary, and which fields should be retained.
Make ambiguous words measurable
Words such as nearby, accessible, high, inside or best often hide the most important modelling decision.
Instead of asking “Which villages are near a clinic?”, specify whether near means five kilometres in straight-line distance, thirty minutes by road, or another model. Likewise, instead of “Join these layers”, identify the target layer, spatial relationship and what should happen when one feature matches several others.
You do not need to encode every GIS operation manually, but you do need to make the intended meaning explicit enough that you can tell whether the proposed operation matches it.
Review the proposed work before confirming it
When Geobble presents work that requires confirmation, inspect the selected inputs, geographic scope, units, assumptions, output and any displayed credit estimate before proceeding. Supported paid processing is not supposed to begin merely because you opened a public link or typed a starter request.
If the proposed operation combines several consequential steps, narrow or split the task, since a smaller request is easier to verify and correct without losing track of which assumption changed the result.
If the first request is unclear
Do not solve ambiguity by adding more adjectives. Identify the missing decision.
If the area is unclear, name the geography.
If the result changes by date, name the period.
If distance is unclear, name the distance model and units.
If a join is ambiguous, name the target and matching rule.
If the output is unclear, state what should be created and what should remain unchanged.
The question is ready when another person could review the proposed operation without having to reconstruct your intent from the final map.
Related content
How to Review Inputs, Units, and Assumptions — next step
Reviewable AI-Assisted GIS — capability
Plain-Language Spatial Analysis — capability