This page explains how we build places from open sources and verify them against official registers, so you can decide how far to trust a place. Every place row includes verified and listed_by, and the response for a single place also includes verified_by.
We build the places for each country from Foursquare OS Places, OpenStreetMap and Overture Maps. When records in different sources describe the same place, we combine them into one place. We find these matches in two passes:
- Name and location. Two records match when their names are similar and they are close to each other. How close depends on the category, so parks and other large sites can be further apart than cafés and restaurants. Records in unrelated category groups, such as a gym and a church, do not match unless one of them has no category or a broad one such as
store. - Location and category. A record that the first pass left unmatched can match a record of the same category at a much shorter distance, whatever their names. For example, this joins a café listed under its trading name in one source and its company name in another. This pass skips crowded locations and broad categories such as
store, where distance alone is not enough.
Registers are official lists, such as food hygiene inspections, health providers and schools. Only records that a register lists as open can verify a place.
A register record verifies a nearby place when the two names identify the same establishment. The names must be similar, and at least half the words in the register's name must appear in the place's name (ignoring words such as the and ltd), unless the two names are the same once spaces are removed, such as "Pizza Express" and "PizzaExpress".
When the two are of different kinds, a word that many places share, such as a town or street name, is not enough to match them. This stops the business next door from being verified. A street or town word that appears in both addresses does not count either when their house numbers differ. Names that pass the similarity test but fail the other checks can still match if they have the same street and house number.
A register that matches by location also matches a nearby place in its category group when every word of one name appears in the other. For a register that only verifies, one of those words must be distinctive, rather than a kind of place or a town.
When a register matches a place, the register is added to the place's verified_by and the place becomes verified.
Some registers give the name of the licensee or company rather than the name over the door, or give no name at all. When one of their records does not match any place by name, it pairs with the nearest place of the same category close by.
Registers that give only a postcode match on the postcode and a stricter name test. So does a record whose location is some distance from the place. It pairs with the only unmatched place in its postcode that passes the stricter test, if that place is within 2 km and does not give a different house number.
Registers can confirm places, but they never delete a place. A register that we use as the primary source for a category also works the other way. A place it lists is included even when no other source lists it, and a place it does not list loses that category.
In GET /v1/catalog/{country}, verified_by_count on the places dataset gives the number of registers a country has. The Coverage page lists the United Kingdom's registers, with their publisher, number of records and match accuracy.
For each register with a judged sample, an AI checks a random sample of its matches and decides whether each pair is the same place. The Coverage page gives the share of pairs judged to be the same place for each register, with a 95% interval.
The table below shows what to send, depending on how strict you need the results to be.
| To get | Send |
|---|---|
| Reliable results for most uses | Nothing. By default, the Places endpoints return only verified places, public transport places and places with 2 or 3 sources. |
| Only places confirmed by a register or by TravelTime transport data | verified=true. Expect gaps where a category or country has no register. |
| More places, including those only one source lists | min_sources=1 on /v1/places. |
| Only places all three sources list, plus verified places | min_sources=3 on /v1/places. |
To apply a stricter rule of your own, read listed_by from each place row and verified_by from a single place.