Methodology
Overview
Every score on this site is computed from published timetable data. This page explains what the score measures, how to read it, and what we deliberately do not measure.
Published by Fusion Party Australia transport policy team.
- Score range
- 0 to 100
- Data source
- PTV GTFS (data.vic.gov.au)
- Scoring dimensions
- Frequency 50%, Coverage 50%
- Coverage area
- Metropolitan Melbourne
- Data vintage
- PTV GTFS processed 2026-07-12
- Suburbs scored
- 788
Master principle
The score is a generous reading of the network
Where a modelling choice is ambiguous, it is resolved in the network's favour. A low score is therefore a floor: the real experience of a STRANDED suburb is at least this bad, and usually worse.
Three choices carry most of this principle:
- The timetable is taken as a promise kept. Cancellations, ghost services, and delays all make real service worse than scored.
- Walking is measured as the crow flies. Freeways, rail corridors, and rivers can only shrink a real catchment relative to the 800m circle, never grow it.
- Driving comparisons use free-flow times. Until peak data lands, every car-competitiveness ratio assumes the driver hits no traffic, which is the best case for the car and the worst case for the argument public transport advocates would make.
For a full accounting of which assumptions cut which direction, see the assumptions register. Every assumption is tagged: one-sided toward the network, one-sided toward the argument (rare and justified), or neutral.
Scope
What this is and what it isn't
It measures what the timetable promises, not what is delivered.
If a bus is scheduled every 30 minutes, we score it as a 30-minute service — regardless of how often it actually arrives. Real punctuality data is a separately planned feature, not a gap we paper over.
It is deliberately biased toward overstating service quality, so low scores are floors.
See the master principle above. When we must choose between two modelling conventions, we pick the one that makes the network look better.
It is not a punctuality tracker, a fare comparison, or a demand forecast.
Fares appear only as cost anchors, sourced and dated. Speed is not scored (frequent beats occasionally fast). Real-time tracking and peak congestion data are pending data access, not omissions we hide.
It is a resource for suburb-level and route-level policy argument.
The worst-20 list, the band distribution, and the route league tables are designed to answer "where should investment go next?" with evidence, not anecdotes.
It is not a resource for lot-level property decisions.
An address lookup tells you what the timetable promises at that point, but it is not a property valuation tool, a commute-time calculator, or a guarantee of future service. The grid cell underlying a single address may change as the population model or GTFS data updates.
Part 02
The Score Bands
Every address and suburb receives a score from 0 to 100, which falls into one of five color-coded bands. A higher score means more frequent services, closer stops, and better connections.
Unit of analysis
Stop to Cell to Suburb
Individual stops
Every train station, tram stop, or bus stop in the PTV GTFS feed is scored out of 100 based on its frequency (how often services arrive) and coverage (where those services can take you), with penalties for weak links.
Address catchments
When you search an address, we look at every scored stop within an 800-metre walk. Your score is your best nearby service (weighted 70%) plus a diversity bonus for having multiple options (weighted 30%).
Suburb scores
We divide the suburb's real Vicmap boundary into 250-metre grid cells, score each cell like an address, and average them weighted by real population data. This means a suburb only scores well if high-quality transit is actually accessible to most of its residents.
Route scores
Alongside suburbs, we also score individual routes: frequency (45%), catchment population (25%), connectivity to other good services (20%), and directness (10%). A route score measures what the route itself does, not what its stops inherit from a good station nearby.