How our scores are calculated
Every score you see is a weighted composite of real, inspectable sub-scores — the exact weights and formulas are below. No score is generated by AI or guesswork.
Suburb overall score
A suburb's overall score is a weighted average of 7 dimension scores, each 0–100. Weights sum to 1.0:
If a suburb is missing a dimension (e.g. no crime data yet), that dimension is excluded and the remaining weights are renormalised — a missing dimension never silently deflates the score.
Safety
Reported incidents are weighted by severity (violent crime and burglary/robbery count far more than minor property crime), converted to a rate per 1,000 residents, and compared against a national urban benchmark. A suburb at the benchmark rate scores around 62; well below it approaches 100.
Affordability
The ratio of average monthly rent to median monthly income. A ratio of 30% or below (the standard affordability threshold) scores 100; above roughly 60% scores near the floor.
Schools, transport & amenities
Each is a diminishing-returns curve over real counts nearby (schools by phase, public-transport nodes, shopping/food/healthcare/community amenities) — the first few matter most, and having many more adds progressively less. Secondary schools are weighted higher than primary in the schools score, reflecting how families plan around them.
Lifestyle
Green space, food/leisure density and community amenities, blended 70/30 with real resident review sentiment once a suburb has reviews. Before any reviews exist, the score rests entirely on the real amenity data.
Property confidence
A directional signal only — never a valuation. It reflects year-on-year price movement where we have it: flat movement scores a neutral ~60, growth lifts it, decline lowers it, capped at ±15% movement.
Complex / estate overall score
The same weighted-composite approach, across 6 dimensions:
Each complex dimension blends an objective feature checklist (boom gate, guards, backup power, and similar) with recency-weighted resident review ratings. Where a complex has fewer than 5 reviews, its review-derived dimensions are shrunk toward a neutral midpoint rather than letting one or two reviews swing the score — the fewer reviews there are, the more the score leans neutral.
Confidence
Every score also carries a confidence tier — Very high, High, Moderate, Limited or Insufficient — based on the real quality of the evidence behind each dimension (how authoritative the source is, how well it geographically matches the suburb, how fresh it is, and how many observations back it). A low-confidence score is always labelled as such rather than presented with false precision.
What we don't do
- We never generate a property sale price or valuation — no licensed sale-price data source exists for any of our 5 countries.
- We never present a low-confidence score with the same certainty as a well-evidenced one.
- Scores are computed from real data, not written or adjusted by an AI model.