Score my next campaign

You are a fundraising analyst. Before I commit budget on a list of target postcodes, score each one against my existing donor profile. Produce two outputs in this directory:

• answer.md — ranked target list, cut-line recommendation, channel suggestions per top-10.
• answer.html — an interactive map of the target postcodes only, coloured by similarity score 0-100, with the recommended cut-line reflected in the legend.

Inputs

• sample_donors.csv — current donor file.
• sample_campaign_targets.csv — postcodes I’m considering for the next push (column: postcode).

Data sources (no auth required)

POA boundary geometry — ABS ArcGIS REST:

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

Pull both the donor postcodes AND the target postcodes (some may be on both lists — that’s expected).

Census 2021 demographics — ABS Data API or general knowledge of Australian postcode profiles, disclosed at the top of answer.md.

The analysis (write to answer.md)

1. Build the reference fingerprint from my donor postcodes: typical ABS profile (age mix, income, owner-occupier %, volunteering, length of residence, voting lean).

2. Score each target — give a 0-100 similarity score using distance across the fingerprint features. Be explicit about the method (population-weighted, z-score distance, etc.) at the top of the answer.

3. Rank the target list. Show the top 10 and bottom 5 with each postcode’s most distinctive feature (e.g. “2030 — close on owner-occupier and age, distant on rental %”).

4. The cut-line. Recommend a score above which I should mail and below which I should drop. Justify with the deck’s Lorenz-curve logic — 28% of postcodes produce 80% of dollars; don’t dilute the campaign.

5. Channel suggestion per top-10. Direct response (mail / paid social) vs community-led (ambassador / event / referral), based on the postcode’s demographic profile.

6. What I’m missing. Up to 5 postcodes NOT on the target list that look very similar to my donor base.

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

• The similarity score (0-100) measures how closely each target postcode matches the average demographics of existing donors — it’s not a prediction of giving.
• Gold-bordered postcodes on the map are above the recommended cut-line — these are the highest-value targets to mail.
• Faded / blue postcodes below the cut-line will dilute the campaign and drag down the response rate.
• Channel suggestions are based on each top-10 postcode’s dominant demographic profile, not on past response data.

Length: 500-700 words plus the explainer. End with the correlational caveat.

The map (write to answer.html)

Choropleth of the target postcodes only, coloured by similarity score, with the cut-line baked into the legend so the audience can see which targets fall above and below.

Required spec

• Library: MapLibre GL JS via CDN.
• Basemap: CARTO light_all raster, no key.
• Geometry: fetch the target postcodes’ polygons from ABS ArcGIS POA.
• Layers:
  • fill — colour each target by similarity score using a diverging ramp:
    • 80-100 → dark accent #f6b651 (gold, full)
    • 65-79 → mid accent #fbcf80 (gold, faded)
    • 50-64 → neutral #c8c8c8 (grey)
    • 30-49 → cool #7ba6e8 (light blue)
    • 0-29 → cold #3a5b8c (dark blue)
  • line — outline weight 1.0 in #1a1a1a at 0.4 opacity.
  • Cut-line ring — for postcodes scoring exactly above the recommended cut-line (e.g. ≥ 65), draw a second line layer at width 2.5 in solid gold #f6b651 so the in-vs-out boundary is visually obvious.
• Initial view: fitBounds to all target polygons.
• Hover popup showing:

  Postcode <NNNN> · <suburb name>
  Similarity score: <NN>/100  →  <ABOVE/BELOW> the cut-line
  Top match feature: <feature> (<value> vs donor avg <value>)
  Top mismatch feature: <feature> (<value> vs donor avg <value>)
  Channel suggestion: <direct response | community-led>
  

• Legend (bottom-left):
  • 5 colour swatches with score bands and labels (“≥80 — strong match”, etc.).
  • A separate gold-bordered swatch labelled *“Above cut-line (≥ ) — mail this"*.
  • A muted swatch labelled “Below cut-line — drop or rework”.
• Title bar (top-left): “Campaign target match scores”. Subtitle: *“Cut-line recommended at — keeps the top 28% of look-alikes, drops the diluting tail."*.
• Explainer panel in the top-right (or under the title bar on narrow viewports) — small card titled “How to read this map”:

  Each target postcode is scored 0-100 for demographic similarity to your existing donors. Gold-bordered postcodes are above the cut-line — mail these. Blue postcodes are below — drop or rework. The score is similarity, not predicted giving. ~300px wide, semi-transparent white, subtle shadow, dismissible.

Constraints

• Self-contained HTML, no build step.
• File size <600 KB.
• Works in Safari and Chrome on macOS.

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 URLs above.
• Don’t claim the ranking predicts giving — only similarity to my existing donor base. The bias caveat applies.
• Be honest about thin signal. If fewer than 200 donors, warn me the fingerprint is noisy.
• Top-10 only if they’re genuinely distinguishable. If the top 4 are statistical twins, say so.
• Be terse. 500-700 words in answer.md.

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.