Methodology

What a 0–100 score means

A score is a neighbourhood’s POSITION relative to other neighbourhoods of THE SAME city, expressed as a percentile. A score of 80 means: “on this criterion this neighbourhood does no worse than 80% of its city’s neighbourhoods, itself included”. The best neighbourhood in a city always scores 100.

A score is not an absolute value. 100 means the best score in that city — not compliance with any standard and not a perfect result. A neighbourhood scoring 100 in a city with poor air may still have worse air than a neighbourhood scoring 40 elsewhere.

MIND THE DIRECTION OF THE SCALE. The “Property prices” criterion is scored for AFFORDABILITY: the lower the median PLN/m², the higher the score. A score of 90 means a cheap neighbourhood, not an expensive one. “Air quality” and “Noise” are inverted the same way — a higher score means cleaner air and less noise.

That is why we do NOT compare scores between cities and do not publish a national ranking of neighbourhoods. The comparison is only meaningful within one city — which is how people actually choose a home.

No data is not zero

Where a criterion has no data for a neighbourhood we show “no data”. We do not substitute zero, estimate a value, or borrow the result from a neighbouring district.

In a combined score a criterion without data is simply left out — the average is taken over the criteria that do have data. Missing data therefore neither drags a score down nor props it up.

If every neighbourhood in a city has an identical value for a criterion, that criterion distinguishes nothing and we show “no data” instead of a score. A ranking in which everyone is equal carries no information — and awarding all of them 100 in that situation would simply be untrue.

Weights and the combined score

Each of the six criteria can be weighted Off, Low, Medium, High or Max. The combined score is the weighted average of the criteria scores, rounded to a whole number.

A criterion set to Off takes no part in the calculation — it does not count as zero.

In the Matchmaker we show how many points each criterion contributed. Those contributions sum exactly to the combined score, so a result can always be decomposed.

How current the data is

Each criterion has its own refresh cycle, following its publisher’s — from measurements updated regularly to registers published once a year. We show the date of the last recomputation with the data, and flag it explicitly when data is stale.

Not every source can be kept current. The air-quality score rests on the European Environment Agency’s validated annual map, and the EEA publishes it roughly two years in arrears — the freshest trustworthy year is not last year. The year the data comes from is stated beside the figure in the report.

A purchased report is frozen: it holds the data as at the moment it was generated and does not change afterwards. The date it was produced is stated in it.

What a score does not tell you

A score describes a neighbourhood, not a particular building or flat. Two addresses on opposite sides of the same street share a neighbourhood score even though they may genuinely differ.

We do not assess a building’s condition, the immediate neighbours, sunlight, how well the building is managed, zoning plans, or developments not yet built.

A score is not advice, a valuation, an appraiser’s opinion or an environmental assessment. We do not analyse any individual’s situation or needs and make no recommendation — for a given address the score is identical whoever asks. It is a starting point for a conversation and a visit — not a replacement for either.

Driving to an emergency department in the report

The report names the hospital emergency department (SOR) fastest to reach BY CAR from exactly the address the report is about. The route runs over the OpenStreetMap road network at free-flow speeds — which is why the time is a lower bound (“from X min”): real traffic takes longer, and no one here has measured a trustworthy congestion multiplier. Candidates are picked by straight-line distance, but the winner by road travel time; in a measured sample of 3,000 Warsaw addresses those differed for 18% of them, which is why we never say “nearest”.

This is NOT the statutory response-time metric of emergency medical teams. That stopwatch (art. 24 of the Act on State Emergency Medical Services) starts when the dispatcher takes the call, and the team departs from wherever it happens to be — it measures an entirely different journey than a resident driving to hospital. Nor is there one direction of travel: a resident drives TO the department (free-flow UNDERSTATES that time), while police or a medical team drive to the address under lights (the same model OVERSTATES theirs) — which is why each service is described separately, in its own direction. For the same reason we never set our figure against the statutory standards: for emergency departments the law defines no travel-time limit at all.

The children’s SOR is shown separately: an adult is not admitted there, so folding both kinds into one “fastest” figure would understate the real drive.

The list of departments comes from the RPRM register, exported at every travel-time recompute; that register as-of date is shown next to the figure. A purchased report captures the data as of the moment it was generated and never changes afterwards. A ward can be reorganised or closed after the register’s as-of date.

Methodology change log

We keep developing the methodology. Changes that affected how scores are calculated are recorded below. A report bought before a change stays frozen in the version it was bought in — see “How current the data is”.

• 07.2026 — “Green & nature”: the score is now based on green and blue land cover (OpenStreetMap). It previously used the GDOŚ nature-protection register, which records legal protection rather than urban greenery.
• 07.2026 — A criterion on which every neighbourhood in a city has an identical value now returns “no data” rather than a score.
• 07.2026 — “Noise”: a neighbourhood lying entirely outside the EEA map coverage now returns “no data”. A missing measurement was previously counted as quiet.
• 07.2026 — “Air quality”: the score is now based on the ANNUAL mean PM2.5 from the EEA’s interpolated 1 km map. It previously used the current reading of the nearest GIOŚ station — that is one day’s weather rather than a property of the place, and Poland has too few stations to tell the districts of one city apart. The report no longer shows an indicative verbal rating against the WHO 24-hour guideline: the score is a position within the city, not a statement that the air is good or bad.
• 08.2026 — The report gained an emergency-department section: the facility fastest to reach by car from exactly the report’s address, over the road network at free-flow speeds, with a separate line for the children’s SOR. Details and limits — the “Driving to an emergency department” section above.