What’s my donor fingerprint?

You are a fundraising analyst. I’ll give you a CSV of donor postcodes and gift amounts. Produce two outputs in this directory:

• answer.md — the demographic profile of my donor file.
• answer.html — an interactive map of my donor postcodes.

Inputs

• sample_donors.csv — columns: donor_id, postcode, gift_amount, frequency, first_gift_date.
• Nothing else is bundled. Pull demographics + boundary geometry from the public ABS endpoints below.

Data sources (no auth required)

Postcode boundary geometry — ABS ArcGIS REST, returns GeoJSON:

https://geo.abs.gov.au/arcgis/rest/services/ASGS2021/POA/FeatureServer/0/query?where=poa_code_2021%20IN%20(<comma-quoted-list>)&outFields=poa_code_2021,poa_name_2021&outSR=4326&f=geojson

Replace <comma-quoted-list> with e.g. '2030','2027','2088' (URL-encode the apostrophes as %27).

Census 2021 demographics at the POA level — ABS Data API (SDMX):

https://data.api.abs.gov.au/rest/data/ABS,C21_T01_POA,1.0.0/...?format=jsondata

The SDMX dimension model is awkward; if you can’t get it cleanly, fall back to your general knowledge of Australian postcode demographics and state your sources at the top of answer.md.

Federal voting (2022 TPP) at booth level — AEC publishes CSVs at https://results.aec.gov.au/27966/Website/Downloads/. Optional — only needed if voting features come up in the analysis.

The analysis (write to answer.md)

Anchor every finding in a real number from the data. Don’t make claims you can’t back. Six sections, in this order:

1. Age tilt — what age bracket are my donors over/under-indexed on vs the Australian average?
2. Housing tilt — owner-occupier / renter / flat-dwelling distribution vs the AU baseline?
3. Income tilt — median household income of donor postcodes vs the national median (~$1,750/week, ABS 2021).
4. Voting tilt — ALP vs LNP TPP lean vs the national 2022 average (52.13% ALP nationally).
5. Community signals — % volunteering, % providing unpaid care vs the AU average.
6. Bottom line in one sentence — “Your donor file is concentrated in [X kind of] postcodes.”

End with three concrete actions (channel / message / segment). Include a final paragraph flagging the correlational-not-causal caveat.

After the actions, add a short “How to read this” block (~80 words) aimed at someone seeing this analysis for the first time. Explain:

• The fingerprint is the average demographic profile across donor postcodes — it’s about places, not individual donors.
• “Over-indexed” / “under-indexed” means how much the donor postcodes differ from the national average.
• The map shows where supporters live (raw count), not where they’re concentrated per capita.
• This fingerprint is the input to every other example in this kit (bias audit, campaign scoring, look-alike modelling).

Target length: 400-600 words plus the explainer.

The map (write to answer.html)

One interactive choropleth of donor postcodes, coloured by donor count.

Required spec

• Library: MapLibre GL JS, loaded from https://unpkg.com/maplibre-gl@4/dist/maplibre-gl.js + matching CSS.
• Basemap: CARTO light_all raster tiles, no key:

  https://{a-d}.basemaps.cartocdn.com/light_all/{z}/{x}/{y}@2x.png
  

  4-domain sharding. Attribution: © OpenStreetMap, © CARTO.

• Geometry source: ABS ArcGIS POA endpoint above, fetched at page load. Cache the GeoJSON in a <script> tag inline if it’s small, or embed the fetched feature collection directly so the file works offline once generated.
• Layers:
  • fill — colour each postcode polygon by donor count using a 5-step Viridis-style ramp (use d3-scale-chromatic or a hard-coded ramp: #440154 → #3b528b → #21918c → #5ec962 → #fde725). Bands at the 20/40/60/80 percentiles of the donor-count distribution.
  • line — thin postcode outlines (#1a1a1a, width 0.5, opacity 0.4).
• Initial view: fitBounds to the bbox of the donor postcodes, padding 40px.
• Hover popup on each postcode showing:

  Postcode <NNNN> · <suburb name>
  <N> donors  ·  $<total amount> total
  Demographics: <age 65+ %>, <owner-occupier %>, <median hhd inc>
  

• Legend in the bottom-left corner showing the colour ramp + donor count breakpoints.
• Title bar in the top-left: “Donor postcodes — coloured by count”.
• Explainer panel in the top-right (or under the title bar on narrow viewports) — a small card titled “How to read this map” with 2-3 short lines:

  Where your supporters live — coloured by raw donor count. Darker = more donors in that postcode. The fingerprint in the analysis is the average demographic profile across these postcodes. Hover a postcode for its demographics. Card width ~280px, semi-transparent white background, subtle shadow, close-button so the user can dismiss it.

Constraints

• The HTML must be self-contained — open the file with open answer.html and it works. No build step. No bundler. Load MapLibre from CDN.
• The map must work in Safari and Chrome on macOS.
• File size budget: <500 KB including the embedded GeoJSON.

When you’re done

After both files are written, run open answer.html so the map opens in the user’s default browser straight away.

Rules

• No web tools beyond the ABS / CARTO / AEC URLs above.
• Cite numbers. Vague qualitative answers fail the test.
• Don’t oversell. End with the correlational-not-causal caveat.
• Be terse. Expected answer.md is 400-600 words.

Disclaimer

Mapulus provides this prompt as educational starter content. The analysis, predictions, and recommendations produced by running it come from a third-party LLM operating on your data — Mapulus is not responsible for the accuracy, completeness, or fitness-for-purpose of any output. Treat the results as hypotheses worth testing, not advice worth acting on without your own validation.