Melbourne Liveability Index
2025 Data and geospatial engineer
PythonNext.jsReactPostgreSQLGeospatial

Melbourne Liveability Index

Interactive web map visualizing and comparing liveability across Greater Melbourne's 673 suburbs across five key dimensions—safety, transport, schools, green space, and housing affordability.

View on GitHub

Key outcomes

  • Computed liveability scoring across 673 Melbourne suburbs.
  • Combined five public datasets into a single explainable index.
  • Delivered an interactive map for suburb comparisons and exploration.

Journal

Started this project because choosing where to live in Melbourne shouldn’t be a guessing game. Real estate websites show prices, but they don’t tell you whether it’s safe, whether you can get to work on public transport, or if your kids would have good schools nearby.

The index combines five critical factors into a single liveability score for all 673 suburbs in Greater Melbourne — safety, transport accessibility, school quality, green space, and housing affordability. Users can search, explore the map, and instantly understand which neighborhoods rank highest across each dimension.

How it works:

The scoring model blends five metrics into a 0–100 liveability score:

  • Safety (25%) — Crime rate per 100k population from Victoria Police data
  • Transport (25%) — Proximity to train/tram/bus stops + service frequency via PTV GTFS
  • Schools (20%) — School quality via ICSEA (Index of Community Socio-Educational Advantage) scores from ACARA
  • Green space (15%) — Parks and accessible green areas from OpenStreetMap
  • Affordability (15%) — Median property prices (inverted so higher prices = lower affordability score)

Each metric is normalized so they contribute equally despite wildly different original scales — crime rates, school indices, and dollar amounts all live on the same 0–100 scale.

The interface:

Light, modern UI with a clean white theme. Search any suburb and the map jumps to it. Click any suburb on the map to see its overall score and full breakdown across all five categories. A sidebar shows detailed metrics and comparisons. The 673 suburb polygons are colored dynamically — cyan for highly liveable, red for less so — so patterns across the city jump out visually.

Tech stack:

Python FastAPI backend serves the scoring logic and data queries. Next.js React frontend handles the interactive map and search. PostgreSQL stores suburb boundaries, scores, and all metric data. Spatial rendering is optimized so dragging and clicking across 673 polygons stays smooth.

Real-world use:

Compare neighborhoods before moving. Evaluate suburb potential for property investment. Or just explore what makes some areas of Melbourne more liveable than others. The data is real — Victoria Police crime stats, ACARA school data, PTV transport schedules, OpenStreetMap parks — so the insights are actionable.

How I used AI

Claude was instrumental in designing the data architecture and helping me think through the weighting model. I used it for scaffolding the FastAPI backend, building the Next.js map components, and optimizing the spatial queries. The product vision — what metrics matter, how to weight them, what the interface should feel like — was mine. AI accelerated the implementation so I could focus on data quality and user experience.