Datadory notebook
Restaurant POI data with chain labels: which sources flag the brand
Restaurant POI data with chain labels is best sourced from Overture Maps Places, whose August 2026 release carries about 74 million POIs - roughly 9.98 million Food and Drink features with a brands attribute and existence confidence score under CDLA Permissive 2.0/Apache 2.0. Foursquare Places API adds realtime chain detection across 100M+ POIs in 200+ countries.
1,744 datasets. Pick your catch.
What chain labels actually are in restaurant POI records?
A chain label ties one point of interest to a parent brand, so every "Starbucks" row rolls up to the same entity instead of sitting in your table as thousands of look-alike coffee shops. In the sources Datadory catalogs for restaurants, the label arrives three different ways.
Overture Maps Places ships an explicit brands attribute on each conflated place record, alongside the name, address, website, phone and a roughly 2,300-category taxonomy; the August 2026 release holds about 74 million POIs, of which roughly 9.98 million are Food and Drink features. The Foursquare Places API treats chain as a query dimension rather than a static field - you search restaurants by category, chain, price and hours across 100M+ POIs in 200+ countries and territories, and every record carries date_created, date_refreshed and date_closed fields that tell you whether a specific franchise location is still trading.
The government feeds take no position on branding at all. NYC DOHMH Restaurant Inspection Results identifies establishments by doing-business-as name within New York City's five boroughs, and UK FHRS Open Data carries business_type plus trade name across roughly 380 local authority feeds - neither publishes a brand identifier, so any chain grouping on those sources has to be derived downstream.
Do government datasets label chains? A UK and NYC comparison
No official inspection feed in this slice publishes a brand identifier, but both major schemes still resolve to named operators, which makes them usable for chain-level hygiene views once you group records yourself.
UK Food Hygiene Rating Scheme (FHRS) Open Data covers every rated food business in England, Wales and Northern Ireland plus FHIS-rated establishments in Scotland, split across roughly 380 local authority feeds refreshed daily through a keyless JSON/XML API requiring no registration. Each record carries the 0-5 rating, RatingDate history, geocoded address and business type; Camden alone publishes 4,260 records (~4 MB of XML). Group those rows by trade name and a national chain's hygiene spread appears council by council - badge reuse follows the FSA terms, with images additionally governed by the FHRS/FHIS Brand Standard.
Where to go next
These pages extend this cluster:
Pick up where this leaves off
Every one of these ships with sample rows before you commit to anything.
Foursquare Places API
Want rows instead of a pitch? Name the datasets.
API, files, or your warehouse. Daily, weekly, or hourly.
Get a sampleQuestions worth asking
Where can I get restaurant POI data with chain labels for free?
Overture Maps Places is the free option: each monthly release holds about 74 million POIs, including roughly 9.98 million Food and Drink features in the August 2026 build, shipped as Parquet under CDLA Permissive 2.0/Apache 2.0. Records carry a brands attribute, a roughly 2,300-category taxonomy, addresses, websites, phones and an existence confidence score.
Can I track hygiene scores for individual chain branches?
Yes. UK FHRS Open Data rates every establishment across roughly 380 local authority feeds with a 0-5 score, inspection date and geocoded address through a keyless daily-refreshed API, and NYC DOHMH Restaurant Inspection Results publishes about 295,000 violation rows with letter grades for all five boroughs. Neither carries a brand field, so aggregate by establishment trade name after cleaning.
How many restaurants are in the Overture Maps places dataset?
About 9.98 million of the roughly 74 million place features in the August 2026 release are Food and Drink category features. The theme ships as tens of GB of Parquet with names, addresses, categories, contact details, brand associations and an existence confidence score, so country-scale restaurant extracts run locally without API calls.