🆕OS NGD Functional Areas
The Functional Areas Collection was added to the OS NGD Administrative and Statistical Units Theme in March 2026. At launch, this collection contains three new retail areas feature types: Retail Area Aggregated, Retail Area Major and Retail Area Minor. Over time, more feature types will be added to the collection that cover other kinds of functional areas, including office areas, leisure areas and more.
Unlike the Boundaries Collection in the same theme, features in the Functional Areas Collection are algorithmically derived from address data, rather than being defined by authoritative organisations or surveyed boundaries.
The attribution across the three retail areas feature types includes geometry, descriptions, retail hierarchy (local, regional, national), retail setting, and address counts. Information on how attribution is processed, integrated and applied is available from the OS NGD Documentation site.
The retail areas feature types have related entity components, which allow them to be connected with the rest of the OS NGD. This ensures functional areas can be analysed and consumed as part of a wider OS NGD ecosystem.
Feature types
Retail Area Aggregated features are derived from areas that contain at least one Retail Area Major feature and any neighbouring addresses with a classification that is indicative of retail use. As a result, Retail Area Aggregated features might be a single parade of shops or a large section of a city. Retail Hierarchy attribution is included on Retail Area Aggregated features to identify if the area is of local, regional or national significance.
Retail Area Major features are identified where addresses with a suitable classification (those which are indicative of retail) share a street name or site and meet the count threshold (as defined by the local address density). Retail Setting attribution is included on Retail Area Major features to give a high-level description of the type of retail setting assigned to the retail area: high street, shopping centre or retail park.
Retail Area Minor features are identified where three or more addresses with a suitable classification (those which are indicative of retail) are within a defined proximity of each other. These features can either be linked to a site or remain unbounded, and they are not constrained by street.
Key features
Algorithmically derived retail geography, based on consistent and repeatable rules.
Three complementary retail layers (minor, major and aggregated), supporting analysis at multiple scales.
Authoritative, standardised definitions of retail areas across Great Britain.
Retail Hierarchy attribution available on Retail Area Aggregated features, identifying areas of local, regional and national significance.
Retail Setting attribution (such as high street, shopping centre and retail park) available on Retail Area Major features.
Integrated with other OS NGD datasets, leveraging related entity components to connect retail area features to other datasets, including OS NGD Address.
Designed for comparison and benchmarking, removing ambiguity in retail geography.
Public-sector ready, supporting policy, planning, evaluation and intervention analysis.
Benefits
Retail area features deliver a consistent, authoritative view of retail locations across Great Britain. They meet a critical need for reliable, comparable data on where retail clusters exist and how they function. Grouping retail addresses into major, minor and aggregated areas enables customers to identify, analyse and benchmark retail geographies at local, regional and national scales.
Ultimately, retail areas data can empower organisations to plan with confidence, measure impact and respond to change, making the datasets an essential tool for anyone who needs to understand and act on the dynamics of retail geography.
Known limitations
Weighted centroids can fall outside a polygon: We have chosen to include weighted centroids for the retail area features to enable a number of use cases where a point is easier to handle. Using a position weighted on the retail addresses makes this a better representation for examples like routing where it increases the accuracy of estimations. However, weighted centroids don't always fall inside the polygon (for example, when the polygon forms a donut or crescent shape). Where this happens, the information about that the polygon inherits from the weighted centroid point is not always valid for the polygon. For example, in Chippenham the centroid falls on a river and so is classed as non-urban, despite the polygon being fully urban. This impacts subsequent calculations, such as which Retail Area Aggregated feature is defined as the town centre. The root cause of a weighted centroid not being in a polygon affects all three retail area feature types; currently, under 2.1% of all features in the three retail area feature types are impacted by this issue.
Temporary retail locations: We do not have information about temporary retail locations. Sometimes these happen to be included because they are located in area which also has permanent retail locations. For example, a market held on a high street would be included in a Retail Area Major feature by virtue of the neighbouring shops; however, an ice cream van in a cliff top car park is unlikely to appear in a retail area feature.
Retail Hierarchy values not as expected in places: In a small number of cases for Retail Area Aggregated features, we believe the published Retail Hierarchy attribute values are demotions. For example, an expected Town Centre could be published with a Retail Hierarchy attribute value of Small Town Centre. We are investigating the cause and any improvements will be made available as part of the regular releases.
Cross reference identifiers in the related entities table are stored as a string: Cross-reference identifiers in the Related Entity Related Component table for retail area features are stored as a string. This means you need to change the character type and potential filter by feature type when trying to join the Unique Property Reference Number (UPRN) to other datasets (for example, the Built Address Feature Type) where they are stored as integers.
FAQs
Frequently asked questions (and answers) on OS NGD Functional Areas can be found under the OS NGD Administrative and Statistical Units Theme FAQs on the OS NGD Documentation site.
Last updated
Was this helpful?