thecolab.ai · Impact for Good — the reasoning
Data too diverse for one person, connected by AI
New Zealand's cost-of-living squeeze is scattered across a dozen official datasets that no one person can hold in their head at once. This is an experiment in using AI to bring them together — and turn them into something a person can actually act on.
NZ poverty is not one number — it is income adequacy, food prices, power bills, rent, and where you happen to live. Each lives in a different place: a Stats NZ release, an MBIE CSV, an MSD rate table, an Otago census file, a supermarket website. Reconciling them by hand is slow, and most people never see them side by side.
Fetching scattered datasets, reconciling units and dates, doing the arithmetic, and joining them into one consistent picture is exactly the kind of broad, patient, multi-source work AI does well — and a human analyst rarely has time for. The machine does the gathering so a person can do the judging.
The result is a set of tools where the income side and the spend side meet: drag a policy lever and watch a household cross — or miss — the basic-needs floor; price a real weekly shop; see rent against the subsidy; and tie it all together in one weekly budget.
That is the "for good": using AI and open data to make the cost-of-living squeeze legible and actionable — for a case worker, an advocate, an analyst, or anyone trying to understand why work no longer guarantees a way out.
The point is the breadth. Each tool here draws on official, public data that normally sits apart:
- Stats NZ — Child Poverty Reduction Act measures, the Food Price Index and electricity sub-indices
- MBIE — Tenancy bond data — real median market rents by district
- MSD / Work and Income — benefit base rates and Accommodation Supplement maxima
- University of Otago — NZDep2023 deprivation by small area, from the 2023 Census
- Woolworths NZ — live supermarket prices for a real weekly basket