An open source delineation and hierarchical classification of UK retail agglomerations
Jacob L. Macdonald; Les Dolega; Alex Singleton (2022). Scientific Data, 9(1). DOI: 10.1038/s41597-022-01556-3
Abstract
Town centres and high streets typically form the social and commercial cores of UK cities and towns, yet, there is no uniform definition of what a town centre or high street is. In this study the spatial delineations of retail agglomerations are generated using open-source data for England, Wales, Scotland and Northern Ireland. The extent and boundaries of these physical retail areas are identified based on the density and connectivity patterns of individual retail units over space. A high resolution hexagonal grid is superimposed over spatial clusters of retail points and a network-based algorithm used to identify mutually exclusive tracts. Agglomerations are then pruned and fine-tuned according to a series of heuristic rules. Our retail agglomerations represent local commerce areas with shopping amenities and are assigned to a hierarchical classification ranking from the largest Regional Centres, Major Town Centres and Town Centres, down to Small Local Centres and Retail Parks. The classification into one of eleven hierarchies is based on a combination of relative rank in the local area and absolute size of retail units within the area. These retail agglomeration boundaries, hierarchical classification and lookups form an open-source spatial data product available for wide use and research implementation.
Extended Summary
This research addresses the absence of uniform definitions for UK town centres and high streets by creating comprehensive spatial boundaries for retail agglomerations across all four UK nations. The methodology combines open-source data from the Valuation Office Agency for England and Wales with OpenStreetMap data for Scotland and Northern Ireland to identify individual retail units including shops, restaurants, cafes, and leisure facilities. The research employs a sophisticated spatial analysis approach using H3 hexagonal grid systems overlaid on retail point clusters, followed by network-based algorithms to identify mutually exclusive retail areas. Each hexagon covers approximately 2000 square metres with a 50-metre radius, enabling precise delineation of retail boundaries based on density and connectivity patterns. The dataset underwent extensive validation through expert stakeholder consultations and ground-truthing exercises, resulting in 6,423 retail agglomerations across the UK: 5,611 in England, 341 in Wales, 392 in Scotland, and 79 in Northern Ireland. The study introduces an eleven-tier hierarchical classification system ranging from Regional Centres serving wide catchment areas to Small Local Centres providing neighbourhood services. This classification considers both absolute retail unit counts and relative rankings within administrative boundaries including regions, ceremonial counties, and local authority districts. Major centres require minimum thresholds of 250 retail units, whilst smaller centres need at least 10 units to qualify for inclusion. The research reveals significant spatial patterns, with London containing approximately 17.5% of all retail agglomerations, predominantly concentrated in Local and Small Local Centre categories. Technical validation against Ordnance Survey High Streets data demonstrates 30-50% spatial overlap, confirming the robustness of the delineation methodology. The comprehensive dataset includes conventional naming conventions derived from street addresses, OS Open Names database, and administrative geography, facilitating practical application in policy and planning contexts. This open-source spatial data product addresses critical gaps in retail geography research and urban planning practice. The systematic approach enables monitoring of retail centre performance, supporting evidence-based policy decisions regarding high street regeneration and retail planning. The dataset provides essential infrastructure for understanding changing consumer behaviours, retail property markets, and economic development patterns across diverse urban and rural contexts throughout the United Kingdom.
Key Findings
- The research identifies 6,423 retail agglomerations across the UK using open-source data and hexagonal grid spatial analysis methodology.
- An eleven-tier hierarchical classification system categorises centres from Regional Centres to Small Local Centres based on size and administrative ranking.
- London contains 17.5% of all UK retail agglomerations, with concentrations in Local and Small Local Centre categories.
- Technical validation shows 30-50% spatial overlap with Ordnance Survey High Streets data, confirming methodological robustness.
- The open-source dataset enables systematic monitoring of retail centre performance and supports evidence-based urban planning decisions.
Citation
@article{macdonald2022open,
author = {Jacob L. Macdonald; Les Dolega; Alex Singleton},
title = {An open source delineation and hierarchical classification of UK retail agglomerations},
journal = {Scientific Data},
year = {2022},
volume = {9(1)},
doi = {10.1038/s41597-022-01556-3}
}