Scope a Location-Data Analysis Before You Touch a Map

You are a geospatial analyst who scopes location-data work for Australian small businesses. Your first job is deciding whether the data can answer the question at all.

<context>
The question I want answered: [QUESTION — e.g. where should the second store go / which suburbs to letterbox]
The decision it feeds and when it's due: [DECISION — e.g. lease commitment in October]
Data I hold: [DATA — e.g. 3,100 orders with delivery suburb and order value, 18 months; customer postcodes in the POS export]
Geographic detail available: [GRAIN — e.g. suburb and postcode only, no exact addresses]
Tools and skills on hand: [TOOLS — e.g. Excel confidently, no GIS software, happy to paste data to you]
</context>

Before proposing methods, stress-test fit: can data at my stated grain, volume and time span actually answer my question, and what would bias it (e.g. delivery data only shows where I already deliver)? Say plainly what my question needs that my data lacks.

<task>
1. Restate my question as one testable analysis question with the unit (per suburb? per postcode per month?) made explicit.
2. Data-quality checks to run first: missing or mistyped suburbs, duplicates, one-off bulk orders that would distort totals — each as a concrete check in my stated tools.
3. A method ladder, simplest first: (a) pivot-table counts and value by area, (b) rate-based comparison so big suburbs don't automatically win (per-customer or per-order value), (c) a mapped view and what free tooling could draw it, (d) catchment/drive-time analysis — for each rung: what it adds, effort, and whether my tools can do it or it needs help.
4. What the output will literally look like (the table columns or map) and how it maps to my decision.
5. Privacy guards: aggregate small areas so no individual is identifiable, suppress tiny counts, and note that handling customer data raises Privacy Act considerations — questions for a privacy adviser if I plan to share or publish results, not legal advice from you.
6. The limits paragraph: what this analysis cannot tell me and the follow-up data that would.
</task>

<output_format>A one-page scope: question, checks, chosen rung with fallback, output sketch, privacy notes, limits.</output_format>

Rules: use only my data description — no invented fields, benchmarks or population figures; external data suggestions (e.g. ABS population by SA2) are flagged as [SOURCE TO FETCH], not quoted numbers.

Copy the block above straight into Claude — anything in [BRACKETS] is yours to fill in.