Every score on this site comes from public data, computed the same way for every suburb in a city. No survey, no editorial opinion, no paid placement. This page states exactly what goes in.
| Safety | Recorded criminal offences per 1,000 residents, from the city's official crime statistics agency, converted to a 0–100 percentile against every other suburb in the same city. |
| Activity | Density of public GPS-activity traces (OpenStreetMap, ODbL) — the same signal behind the "Strava heatmap" trick, but blended with the signals below rather than used alone. |
| Wealth & Value | Median household income vs median rent (ABS Census 2021). Value is left unscored where a suburb's rental market is too thin to trust. |
| Parks & Green space | OpenStreetMap park polygons — hectares per 1,000 residents and share of the suburb within walking distance of a park. |
| Cafés & Transit | OpenStreetMap points of interest — density per resident and per km², so a big suburb cannot outrank a small one purely on raw counts. |
| Transit | OpenStreetMap public-transport stops inside the suburb boundary or within 100 m of it (a stop on the boundary street serves both sides) — train stations, light-rail and tram stops, bus stops and ferry wharves, per km². A station counts as 5 stops and a ferry wharf as 2, because one station carries far more trips than one bus stop. Stops are counted, not timetables: frequency is not yet measured. |
| Parkrun & Healthy/active | Distance to the nearest parkrun event, plus gym/sports-facility density per resident. |
| Melbourne | Crime Statistics Agency Victoria |
| Sydney | NSW Bureau of Crime Statistics and Research (BOCSAR) |
| Brisbane | Queensland Police Service |
| Perth | no suburb-level crime source |
| Adelaide | South Australia Police (SAPOL) |
| Melbourne | Sale prices: Victorian Property Sales Report, October–December 2025 quarter — © State of Victoria (Department of Transport and Planning / Valuer-General Victoria), Victorian Property Sales Report, CC BY 4.0. Quarterly median of sales settled in the quarter, per suburb. Figures marked as low-volume rest on fewer than 10 sales that quarter. |
| Sydney | Sale prices: NSW DCJ Rent and Sales Report, January - March 2026 quarter — © State of New South Wales (Department of Communities and Justice), Rent and Sales Report, CC BY 4.0; suburb-to-postcode mapping © GeoNames.org, CC BY 4.0. NSW publishes no free suburb-level sale medians, so this is the median for the suburb's POSTCODE (non-strata sales = houses, strata = units), which can span several suburbs. Low-volume figures rest on 30 or fewer sales in the quarter; postcodes with 10 or fewer sales are withheld by DCJ. |
| Brisbane | Sale prices: Queensland Regional Profiles (QGSO), 12 months ending 31 March 2026 — © The State of Queensland (Queensland Treasury), Queensland Government Statistician's Office, Queensland Regional Profiles, CC BY 4.0; suburb-to-SA2 placement via ABS ASGS 2021 (CC BY 4.0). Queensland publishes sale medians per ABS Statistical Area Level 2 (SA2 — "broadly equivalent to suburb size"), not per suburb, so this is the median for the SA2 the suburb sits in (named next to each figure when it differs); detached dwellings = houses, attached = units. The window is the latest 12 months, not a quarter, and QGSO withholds medians under 10 sales. |
| Perth | Median sale price: not available for Western Australia — Landgate sells suburb sales reports and REIWA's medians are commercial; there is no open-licence suburb-level sale price dataset for WA. |
| Adelaide | Sale prices: SA Government Metropolitan Median House Sales, April–June 2026 quarter (2Q 2026) — © Government of South Australia (Land Services SA), Metropolitan Median House Sales, CC BY 4.0. Quarterly median house sale price per suburb, metropolitan Adelaide. Units are not in this dataset. Figures marked as low-volume rest on fewer than 10 sales that quarter. |
| Safety | 30% — recorded offences per 1,000 residents (percentile). Omitted, and the rest renormalised, where a state publishes no per-suburb crime data. |
| Schools | 25% — half is schools per 1,000 residents, a quarter is a primary school within 800 m of the suburb centre, a quarter a secondary within 2 km (linear fall-off beyond). Counts come from government school-location datasets where the licence allows (Victoria all sectors; NSW, Queensland and SA government schools only), otherwise OpenStreetMap. School quality (My School / VCE) is deliberately absent: those results may not be republished per school. |
| Parks & green space | 15% — parks per 1,000 residents and share of the suburb within 400 m of a park. v1 used park hectares, which penalised dense inner suburbs like Surrey Hills for being built-up rather than for lacking parks. |
| Kinders & childcare | 10% — OpenStreetMap kindergartens and childcare per 1,000 residents. |
| Price for the size | 10% — median house price against the city median, gently: full marks at 0.8× or below, half at 1.2×, none at 1.6× or above. Omitted where no open sale-price data exists (Brisbane, Perth). |
| Transit | 10% — the transit score. |
Why it changed: under v1 (safety, parks, green space, OpenStreetMap school points at half weight) suburbs such as Glen Waverley and Surrey Hills ranked poorly — partly a data bug (points outside the city bounding box were never fetched, so Glen Waverley showed one school and four cafés; fixed the same day) and partly a formula that rewarded hectares of parkland over what families actually look for. v2 states each weight, and every suburb page shows its inputs.
We do not hold school catchments or school results, so the "for families" ranking counts schools and their distance but not zoning or quality — every families page says so explicitly. Queensland publishes crime data only at police-division level, not per suburb, so Brisbane-area suburbs in the same division share a figure. ABS Census 2021 gives us population, age, income and rent per suburb — not renter/owner split, family composition or birthplace. Median sale prices are medians from state government reports where an open licence exists (Melbourne and Adelaide per suburb and quarter; Sydney per postcode; Brisbane per ABS SA2 over the latest 12 months, from the Queensland Regional Profiles); they are not available for Perth. Suburbs under roughly 1,500 residents are excluded from scoring as too small to trust statistically.
Up to ten 0–100 percentile scores per suburb: safety (official crime statistics), activity (open GPS-activity traces), wealth and value (ABS census household income vs rent), parks and green space (OpenStreetMap park polygons), cafés and transit (OpenStreetMap points of interest), parkrun proximity, and a healthy/active composite. A score is left blank rather than guessed at when the underlying data is too thin to trust for that suburb.
Each city's own police or justice statistics agency — see the per-city table below. Every figure is recorded offences per 1,000 residents, latest published period. Where a state (Queensland) publishes only police-division data, not suburb-level, that is stated on every affected suburb page.
ABS Census 2021, at the suburb (SAL) level, for population, median age, median household income and median weekly rent. We do not hold renter/owner split, family composition or birthplace data for any suburb — those fields are not in our census data, and we say so on every suburb page rather than omitting them silently.
Only free, official datasets whose licence allows republishing with attribution (all CC BY 4.0): the Victorian Property Sales Report (Valuer-General Victoria) for Melbourne, the NSW Rent and Sales Report (DCJ) for Sydney — postcode-level, because NSW publishes no free suburb-level sale medians, mapped to suburbs with GeoNames — and the SA Government's Metropolitan Median House Sales (Land Services SA) for Adelaide; and, for Brisbane, the Queensland Government Statistician's Office's Queensland Regional Profiles (residential dwelling sales from the QVAS database, detached = houses, attached = units), published per ABS SA2 rather than per suburb and as a 12-month window, so each Brisbane figure names the SA2 it belongs to. Western Australia publishes no open suburb-level sale price data (it is sold by Landgate, and REIWA's figures are copyright), so Perth pages say 'not available' rather than quoting a commercial site. Each figure carries its period, and low-volume figures are flagged.
In Melbourne and Adelaide, where we hold suburb-level open sale medians (Victorian Property Sales Report; SA Government Metropolitan Median House Sales, both CC BY 4.0), the first-home-buyer page starts from median house sale price: only suburbs with at least 10 house sales in the quarter are used, the cheapest 40% are kept, any in the bottom quarter of the city for safety or public transport are dropped, and the rest are ordered by an equal blend of price percentile (cheaper ranks higher), safety and transport, top 10. Sydney (postcode-level prices), Brisbane (SA2-level) and Perth (no open sale data) still shortlist by ABS census median rent, and those pages say so.
The site is rebuilt and redeployed as new source data becomes available; each page states its own last-updated date. Census figures are fixed to the 2021 ABS Census until the next release.
No. No suburb, agent or developer can pay for placement, and none has. The formula and inputs are published on this page and reproduced on every city and suburb page.
GeoNames.org's Australian postcode file (CC BY 4.0), matched to each suburb by name and, where a name is shared by more than one postcode, by nearest distance to the suburb's own centre. It does not feed any score.
Updated October 2026 (last updated 2026-10-17). ← back to BestSuburbs