Method and sources

How Home Clue calculates each score, and where the numbers come from. Everything is based on open data you can check yourself.

Ground rules

  • Scores are calculated with fixed, versioned rules. The same data always gives the same score, and no AI sets or adjusts them.
  • When data is missing we show “unknown”. We never fill in an average or a guess.
  • We never assess addresses or areas by who lives there, and we hold no such data. Scores are based only on services, environment, transport and price.
  • Matching against your wishes is a check against your own list, using the same data as the report. Missing data shows as unknown, and it is never advice to buy.
  • If you describe your dream house in your own words, the text is kept in your account until you empty the box and save. An AI service (OpenAI) reads wishes from it and may only pick from fixed kinds of wishes. Places you travel to (work, school, gym …) are looked up on the map; travel time is checked for the first three. The AI never sets a score and never decides whether a house meets a wish – fixed rules do. A text about who lives in an area is never passed on.
  • Street suggestions use the same data and rules as the reports, for several points along each street. A street is judged by its typical point, so one good end doesn’t lift the whole street, and unknown data never counts as met.
  • A report is a snapshot with a source and fetch date for every number. It never changes afterwards; an update becomes a new report.
  • Distances and walking times are estimates, not actual routes. That’s why we write “about”.

Price & value

Sources: Statistics Sweden (SCB), Your input

We estimate a market value from the home’s taxeringsvärde and how villas in the kommun actually sell relative to their taxeringsvärden. An asking price below the estimated value gives a higher score, a price above it a lower one.

Limitations: The comparison uses an average for the whole kommun, so the home’s condition and size don’t show. Without asking price and taxeringsvärde, price shows as unknown. Price level by area (e.g. district) isn’t available yet, as we have no source for it.

Safety

Sources: Swedish Police, Statistics Sweden (SCB)

We compare the crime events the Police publish for the kommun with the other kommuner in the county, relative to population. More serious crimes weigh more. Fewer events than the other kommuner gives a higher grade; a kommun in the middle of the county gets a middling grade, and no kommun gets a full score – a ranking is never a promise of safety.

Limitations: The Police only give positions at kommun level, so the score is the same for every address in the kommun. The events are a selection, not complete crime statistics.

Transport & access

Sources: Trafiklab (SL)

We look at stops within walking distance and how often they are served, especially at peak times. A short walk to frequent departures, particularly rail, gives a higher score.

Limitations: The timetable shows planned service, not delays or crowding.

Schools & family

Sources: Skolverket, OpenStreetMap, Kolada

We weigh how close the nearest grundskola and preschool are and how nearby schools perform according to Skolverket’s statistics. Close schools with good results give a higher score.

Limitations: If school results are missing, the score is based on distance only and marked as partial. Catchment areas and queues are not included. Preschools come from OpenStreetMap, which lacks many preschools in some kommuner. Where the map has too few compared with official statistics, preschool isn’t counted and its distance shows as unknown.

Noise & environment

Sources: OpenStreetMap

Noise is estimated from the distance to major roads and railways. The closer and the bigger the route, the lower the score.

Limitations: This is an estimate, not a measurement. Roads and rail in tunnels are not counted. Noise barriers, terrain and aircraft noise are not included.

Energy & running costs

Sources: Boverket

The score is based on the energy class in the home’s energy performance certificate. A better energy class gives a higher score.

Limitations: Not every home has a certificate; energy then shows as unknown. A certificate can be several years old.

Sources

Swedish Police
The Swedish Police’s published events (polisen.se). A selection of recent events, not complete crime statistics.
Statistics Sweden (SCB)
Statistics Sweden’s database: villa sale prices and population per kommun.
Skolverket
The Swedish National Agency for Education’s school register and statistics: schools, locations, merit scores and other results.
OpenStreetMap
OpenStreetMap: roads, rail, preschools, grocery stores, parks, forest, water, playgrounds, sports grounds, place names and kommun boundaries. © OpenStreetMap contributors.
Kolada
Kolada (RKA): the number of preschool units per kommun from Skolverket’s statistics. Used only to judge how complete the map’s preschools are.
Trafiklab (SL)
SL timetables (GTFS Regional) via Trafiklab: stops and departures.
Trafiklab ResRobot
Travel time to your places: Trafiklab’s ResRobot route planner (Samtrafiken), looked up when you open a report and kept for 30 days.
Boverket
Boverket’s register of energy performance certificates.
Your input
Asking price and taxeringsvärde that you enter yourself. They are stored only in your report.