The whole method, in the open: how nearby places become seven dimension scores and one overall 0–5★ rating. No hidden weights.
For any point you pick, BuildingsScore runs four steps:
Each dimension is scored 0–5 on its own, then mixed into the overall rating with the weight below. The weights reflect how much each tends to shape day-to-day life.
| Dimension | What it measures | Blend weight |
|---|---|---|
| Convenience | Shops, groceries, pharmacies, services within walking distance | 20% |
| Quiet | Freedom from busy roads, rail, nightlife, stadiums, flight noise | 18% |
| Transit | Access to metro, train, tram, bus, ferry | 15% |
| Environment | Distance from industry, waste, farmland, plus flood risk | 13% |
| Nature | Parks, forests, water, playgrounds, sport & recreation | 12% |
| Safety | Emergency services nearby, minus a few negatives | 12% |
| Air quality | A live air-quality index (where available) | 10% |
If a dimension has no data for a spot — most often air quality — it's dropped and the remaining weights are re-normalized, so the overall rating is always a fair average of what we actually know.
Every nearby thing has a range (in metres) beyond which it no longer counts. Within that range its effect fades with distance — full strength right on top, zero at the edge:
When several of the same thing are nearby, the nearest one dominates and extra ones add progressively less (the score saturates) — three supermarkets aren't three times one. Continuous features like a road or railway are treated as a single thing at their nearest point, not counted once per mapped segment.
Different kinds of dimension call for different math:
More good stuff nearby pushes the score up toward 5, with diminishing returns. Each place adds its importance × closeness; the total is mapped onto 0–5 along a saturating curve, so going from nothing to a few amenities matters far more than going from many to many-plus-one.
Transport modes substitute for one another: an excellent metro stop already makes an area well-connected, and lacking a tram doesn't drag it down. So Transit rewards the best access you have rather than summing modes — a town without a subway isn't penalized for it.
These start clean at 5 and each nearby bad thing multiplies the score down by its severity × closeness. One loud neighbour can dent it; several compound. Severity is per-factor — a motorway or landfill weighs much more than a single bar.
Safety starts from a neutral baseline and is lifted by nearby emergency services (police, fire, hospital) and dragged down by a few negatives (e.g. a prison), then clamped to 0–5.
Note: BuildingsScore does not currently use point-level crime data (there's no free, global source — see Data sources). Safety here is a proxy from emergency-service proximity, not a crime statistic.
Where available, a live Universal Air Quality Index (0–100) is mapped directly onto 0–5 stars. Where it isn't, the dimension is simply omitted from the blend.
Some things can't be read off a map of places, so BuildingsScore uses honest approximations:
The final number is the weighted average of whatever dimensions are present, rounded to one decimal — a clean 0.0 to 5.0 in 0.1 steps. Open any dimension in the app to see the individual places driving it, each with its distance and a walking route.
The weights and ranges are opinionated, transparent defaults — a sensible general view of livability, not an objective truth. Your priorities may differ (a student might love the nightlife that costs a Quiet point). Treat the score as a fast, comparable first read, and always visit in person before making a decision.