All member projects

Code · Built by members

NZ Suburb Intelligence

Ask plain-English questions about Auckland suburbs and get cited answers, maps and side-by-side comparisons from free public data.

Conceptual illustration: Ask plain-English questions about Auckland suburbs and get cited answers, maps and side-by-side comparisons from free public data.
AI-generated editorial illustration, not a product screenshot.

What it is

NZ Suburb Intelligence is a suburb comparison tool built on New Zealand open government data, including LINZ, Stats NZ Census 2023 and Ministry of Education data. You type a question, such as which suburbs have the lowest rent, and get a streaming answer where every figure shows its source and date.

It covers Auckland for now and is designed so other regions can be added. It is aimed at people choosing where to live who want facts about housing, people, schools and commutes in one place.

What it does

  • Turns plain-English questions into structured queries and cited answers, and shows how it read the question.
  • Shades a map of Auckland suburbs by any metric, with hover details and fly-to on selection.
  • Shows suburb profiles with census trends from 2013 to 2023, nearby schools and typical travel times to the CBD and airport.
  • Answers commute questions by looking up NZ addresses and routing on real roads, with no live traffic.
  • Compares up to three suburbs side by side and flags which are under or over a weekly rent budget.

Built with

  • TypeScript
  • Next.js
  • Supabase
  • PostGIS
  • pgvector
  • Claude API
  • Gemini embeddings
  • MapLibre GL
  • openrouteservice
  • Vercel
  • civic & data

More from the community

Website

PHP2Shell NZ tracker

by Adam

An hourly tracker of NZ sites and their PHP versions during the PHP2Shell bug, with a free site checker.

Website

Data stories

by Caleb

A growing collection of interactive data stories, with maps and charts on Indonesia, Southeast Asia and New Zealand.

Website

Did they answer?

by Matt

A civic experiment that uses a small AI model to read ministers' replies to written questions in Parliament and say whether they gave what was asked.

Back to all member projects