How we score streets and areas

A computer scores every street the same way, from open data anyone can check. The main score uses no opinions or surveys. Only the separate safety score adds crowd maps, surveys and police records.

1. 1 We map every street Each street is cut into pieces about 60 m long, so one street can score differently at each end.
2. 2 We look from above Summer satellite photos show where there are trees, grass and paving.
3. 3 We measure the view Every 15 m we look across the street: how far away the house opposite is, and how much sky it hides from a first-floor window.
4. 4 We walk it On the real street network we count the minutes to the nearest tram or bus stop, shop, café, pharmacy, doctor, school, gym and park.
5. 5 We listen Big roads and railways nearby make a street loud. Buildings in between soften it.
6. 6 We score it Each part gets 0 to 100 against published guidelines where they exist (WHO noise limits, daylight rules, the 3-30-300 tree rule). One overall score combines them. Some things can't be made up for: a street in a tourist hotspot, on a loud main road, or where you'd need a car can't score high, however nice the rest is.

How we rank areas for families, walking and safety

Each district's score is its typical street: half its streets score higher, half lower.

We also score every area for families (schools and playgrounds close by, parks, quiet, calm traffic), for getting around on foot, and for safety. These help you rank areas. They never change the map colours or the main score. We don't rate how good the schools are.

Safety combines our own street estimate (traffic speed, crossings, lighting, places open nearby) with police records where a city publishes them, the Hoodmaps crowd map, and crime figures for the whole region. Crowd maps are often blunt or joking and nobody checks them, so every source has an (i) note saying where it comes from and how far to trust it.

The data we use

Satellite photos Summer images from the European Space Agency's Sentinel-2 satellites, and their land-cover map.

OpenStreetMap The free world map drawn by volunteers: streets, tram and bus stops, shops and cafés.

Overture Maps Open data on buildings, how tall they are, and the places inside them.

Your city's open data Tree lists, noise maps or 3D buildings, when a city publishes them under an open licence.

Safety sources Police records where a city publishes them, Eurostat crime figures for the region and the Hoodmaps crowd map. Each has an (i) note on the map.

No Google, Apple or Bing data, and nothing about the people who live there.

Where it can be wrong

• Noise is estimated from roads and railways, not measured. Aircraft, bars and neighbours aren't in it.
• Single street trees can be missed; the satellite sees 10 m squares.
• We only look out of street-facing windows, not into courtyards.
• Where building heights are unknown, we estimate them from nearby buildings. That's most buildings in Dnipro and Ljubljana.
• Bus frequencies are assumed per type of transport until we add timetables.
• Shops and services are only as complete as OpenStreetMap and Overture.
• Walking times ignore hills.
• Tourists are a relative score from sights, photos and hotels, not a visitor count.
• Safety mixes our estimate with crowd maps and police records. Crowd maps can be unreliable, and police records miss crimes nobody reports.
• We don't cover prices, rents or school quality.

See something wrong? Click the street on the map and tell us under “Does this sound right?”.

Licences and credit

City data is shared under the Open Database License (ODbL), like OpenStreetMap. The mapping skill is free on GitHub for non-commercial use (PolyForm Noncommercial 1.0.0). Map data © OpenStreetMap contributors and Overture Maps Foundation. Contains modified Copernicus Sentinel data. ESA WorldCover 2021, CC BY 4.0.

Scores are estimates. The site and its data come with no warranty. Please don't rent or buy a home on the score alone.

Privacy

If you click Allow on the banner, we count visits and clicks with Google Analytics: which pages and cities people open, what they search for, and which buttons they use. Nothing is sent to Google before you click Allow, not even for fonts: the site serves its own. We never send your notes or your email address, and we take out anything that looks like one.

The site keeps no personal data about you, except what you type into it: the email address on the business form, notes you write about a street, and the name you give for credit when you add a city.

Like every website, ours sees your internet address when you visit. It uses it only to limit how often one address can send things, and forgets it within an hour. When an AI agent sends a city, we keep the address it came from while the upload waits for review, and delete it a day after that, or after 30 days at most. Our host's request logs also hold addresses for a short time.

To change your mind, use Cookie choices at the bottom of any page.