Method · Hausfuchs

METHOD

How every score is made

Thirteen neighbourhood scores for every street block in Finland. This page says what each one measures, what goes into it and what it cannot tell you.

How to read a score

Every score runs from 0 to 100 for a street block.
Every score is made the same way for every street block in Finland, from the same national data.
For ten scores, the number shows where the block stands among Finnish homes. A 70 means better than about 70 out of 100 homes. It does not mean 70 percent of a maximum. Water and Air quality show the plain score, and Quietness is tied to decibel levels.
A score is left empty when there is not enough evidence behind it. An empty score shows as “No data”, never as 0.
Two cases look alike but are different. When we do not know a value, we leave it out and the rest of the score still counts. When we know there is nothing within reach, for example no pharmacy nearby, that counts as far away and the score is low.
Where an input is only known for a larger area, such as a postcode, the score says so. Not every score has the same precision.

The thirteen scores

Walkability

How easily everyday needs and a pleasant street are reachable on foot.

Inputs: Nearest grocery, pharmacy, restaurant, bakery and shops; walking and cycling paths, street lighting and calm streets.
How it is built: Each input becomes a closeness value that falls with distance. Most of the weight is on destinations and the rest on comfort. Nothing within reach counts as far.
Limits: Straight-line distance. Shops and restaurants partly come from OpenStreetMap. Path and lighting data are missing in some areas, and then they are left out.
How the number is shown: As a position among Finnish homes.

Transit

How close and how frequent public transport is, including rail.

Inputs: Stops, weekday departures and train stations from the national timetable feeds.
How it is built: Distance to the nearest stop weighs most, then how often service runs, then distance to a station. No service within walking reach scores zero.
Limits: No travel-time modelling. Frequency is a weekday average.
How the number is shown: As a position among Finnish homes.

Green & nature

Tree cover, parks, forest and natural land around the block.

Inputs: Forest and tree-canopy data, parks and protected areas from the topographic database, and garden and allotment points.
How it is built: Each part takes the better of an urban-green value and a natural-land value, so a village inside a forest is not penalised for lacking a mapped park.
Limits: Single street trees and garden trees can be under-counted. Gardens and allotments partly come from OpenStreetMap.
How the number is shown: As a position among Finnish homes.

Water

Closeness to sea, lakes and rivers, with water quality and swimming access as extras.

Inputs: Water bodies from the topographic database, ecological status of water bodies, and swimming and beach sites.
How it is built: Closeness weighs most. Water quality counts only where a rated water body is near. Swimming counts the nearest site or open water at the door.
Limits: Only rated water bodies have a quality value. Small ponds are left out. Shown as the plain score, not as a rank.
How the number is shown: As the plain score.

Quietness

How little road, rail, tram, metro and aircraft noise reaches the most exposed facade of a building.

Inputs: A noise model fed by traffic counts and road classes, train traffic, tram and metro timetables, and aircraft noise contours.
How it is built: Noise from each source is turned into expected annoyance and sleep disturbance using the EU dose-response curves, then combined across sources. It is modelled, not measured.
Limits: Streets without traffic counts use typical values for their road class. Industrial and nightlife noise are not included. Aircraft noise covers Helsinki-Vantaa only.
How the number is shown: As the plain score.

Air quality

Long-run outdoor air quality where people live.

Inputs: Annual European air quality maps for nitrogen dioxide, fine particles and ozone, on a 1 km grid.
How it is built: Each pollutant becomes a sub-score, and the blend gives most weight to fine particles and nitrogen dioxide. It is not the EU limit-value scale, which would read near 100 almost everywhere.
Limits: A 1 km grid cannot see a single busy street. Where there is no coverage the score stays empty. Shown as the plain score, not as a rank.
How the number is shown: As the plain score.

Family

How well the surroundings serve families with children.

Inputs: Daycares, primary and secondary schools, playgrounds, and the share of households with children.
How it is built: Distance to each kind of place is blended with fixed weights. Primary school weighs more than secondary school. Education registers are used only to compute the score and are not republished.
Limits: We count places, not school quality or catchment rules. The household share is known per postcode.
How the number is shown: As a position among Finnish homes.

Health & wellness

Closeness to pharmacies, health centres, sports facilities and swimming venues.

Inputs: Pharmacies and health centres from the national amenity layer, and sports and swimming facilities from the national sports facility register.
How it is built: Four distances are blended with fixed weights. A remote home scores low and is not left empty.
Limits: Proximity only, not quality, opening hours or care capacity.
How the number is shown: As a position among Finnish homes.

Economic vitality

How economically active and prosperous the area is.

Inputs: Postcode income, education and jobs, population density, shops, restaurants and the density of registered companies.
How it is built: A weighted blend. Where a postcode hides its income for privacy, the score uses a regional estimate, mostly weighted towards local shops, restaurants and companies. This is the one score where a missing value is filled by an estimate.
Limits: The income side is known per postcode, not per block.
How the number is shown: As a position among Finnish homes.

Building character

How consistent the built form of a block is, in building height and footprint size.

Inputs: Building footprints from the topographic database and building heights from laser scanning.
How it is built: Half from how similar the heights are and half from how many footprints share a typical size. It rewards a consistent street of any kind, not beauty.
Limits: A block needs enough buildings with known heights, so many blocks stay empty. Tree crowns can make some buildings read too tall.
How the number is shown: As a position among Finnish homes.

Road safety & stability

How safe the roads are and how stable the neighbourhood is. It is not a crime measure.

Inputs: Home ownership, unemployment and vacancy per postcode; injury road accidents over ten years; street lighting; and unprotected railway level crossings.
How it is built: A weighted blend where road accident risk weighs most. Accident risk is estimated for each block and pulled towards what its roads would predict, so a few accidents do not swing it.
Limits: No crime data is used. Three of the inputs are known per postcode. This score needs a lower share of evidence than the others, and shows its evidence share.
How the number is shown: As a position among Finnish homes.

Culture & leisure

Closeness to libraries, museums, cinemas, theatres and restaurants.

Inputs: Libraries from the national library register; museums, cinemas, theatres and restaurants from the national amenity layer.
How it is built: Four distances and a restaurant density are blended with fixed weights. A remote home scores from what can be measured and is not left empty.
Limits: Proximity only, not programme, size or opening hours. Several inputs come from OpenStreetMap.
How the number is shown: As a position among Finnish homes.

Terrain

How flat, low and flood-safe the ground is, with a small bonus for south-facing slopes.

Inputs: A 2 m elevation model for slope, height and aspect, and river flood hazard maps combined with how low the ground lies near water.
How it is built: Flatness weighs most, then flood resilience, then elevation and aspect.
Limits: Flood maps cover the major rivers, and a nearness-to-water estimate covers smaller streams and the coast. If any terrain value is missing the score stays empty.
How the number is shown: As a position among Finnish homes.

Modelled, not measured

Noise, sea view and the Helsinki region sun figures are computed with models. They are estimates, not measurements taken at the window. For noise, the national noise maps from the authorities remain the official source, and we link to them rather than replace them.

Beyond the thirteen scores

Three modelled engines work at building level, in the Helsinki region.

Sun. How much direct sun a facade, balcony or roof gets through the year, given the buildings and trees around it. Helsinki region.
Noise by floor. Modelled road, rail and tram noise at the facade, by floor. Helsinki region.
Sea view. How much open sea a window sees. Helsinki region. This is an early version and is not yet calibrated.

What we deliberately do not score

Crime. No crime data goes into any score. Finland has no open crime data at neighbourhood level, so the Road safety & stability score uses road safety and neighbourhood stability only.
Who lives there. No score describes residents.
School or care quality. We count what is within reach, not how good it is.
Prices. Home prices are kept out of every score.
Beauty. Building character rewards a consistent street, not a taste.

Why our number can differ from another source

Our noise score is modelled for the most exposed facade of a building and combines road, rail, tram and aircraft noise. A municipal noise map may use different sources, heights or years.
Our sun figures are modelled from buildings and trees in the Helsinki region. A weather station reports sunshine hours for the open sky, which is a different quantity.

Limits that apply to all scores

Distances are straight-line, not walking routes.
Some inputs come from OpenStreetMap, whose completeness varies by place.
Some inputs are known only per postcode or per 1 km grid cell, and are applied to the blocks inside.
Scores describe a place, not a person. They are decision support, not a valuation.

The full calibration is shared with pilot partners under contract.

Where the data comes from