Investigating retail space performance through spatial configuration of consumer movement: A comparison of York and Leeds

Adejimi Adebayo, Paul Greenhalgh, Kevin Muldoon-Smith

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


Spatial layouts help to shape retail consumer movement, which in turn plays a role in determining the distribution of retailers and performance of retail space on city network. Spatial configuration can be understood through street segment analysis, computing to-movement (integration) and through-movement (choice) metrics within a given set of connecting street networks, making it possible to assign syntactic values to individual street segments (space). In this paper, such syntactic values for the cities of Leeds and York have been established to indicate a spatial accessibility index that can be used to understand potential human (consumer) movement on spatial layouts. Other studies have established relationships between computed syntactic values and ranges of socio-economic activities, including land uses and urban value distributions. However, little is known about how configured (movement) metric outputs relate to changes in retail space’s rental values (as proxy for retail space performance) across different city network scales. In response, this study investigates the relationship between retail space performance and consumer movement patterns (CMP) within sampled spatial layouts. The CMP are defined as spatial configuration metric outputs of integration, choice and normalised angular choice (NACH) metrics, computed at macro (city) and meso (city centre) scales. Street segment analysis on spatial layouts at city (macro) and city-centre (meso) scales were computed using DepthMapX tool to obtain the CMP variables. The computed syntactic values of CMP variables were then exported as point features into QGIS for analysis with the retail space performance within the sampled spatial layouts. Rental value data for years 2010 and 2017 were obtained from the Valuation Office Agency VOA datasets for York and Leeds. The two datasets were linked through a common key variable (Unique Address Reference Number) to compute rental value changes using MS Access and MS Excel tools. The rental value change table was also exported as point features into QGIS for geospatial analysis with the computed syntactic values of CMP variables. To achieve this, the study utilises vector grid (developed at 500m X 500m at city scale, and 200m X 200m at city centre scale for both cities) to a create uniform platform for all variables per grid. The relationship outputs between variables were investigated at macro city scale and meso city-centre scale for the two cities. The study reveals that there are variations in relationships between retail space performance and computed movement syntax across different scales of spatial layouts. The variables exhibit significant positive relationships at mesoscale (city centre), while variables exhibit weak correlation at the macroscale (city) for both cities. It further reveals that the integration (to-movement) metric has the most significant impact on retail space performance, with the through-movement metric having the least impact across all spatial layouts. On this basis, the study conclude that integration metric has the capability of signalling future of retail space (rental value) performance at city mesoscale layouts.

Original languageEnglish
Title of host publicationProceedings of the 12th Space Syntax Symposium (12SSS)
Subtitle of host publication 8-13 July 2019, Beijing, China
Place of PublicationBeijing
PublisherSpace Syntax Network
Number of pages19
ISBN (Print)9781510893795
Publication statusPublished - 8 Jul 2019
Event12th International Space Syntax Symposium, SSS 2019 - Beijing, China
Duration: 8 Jul 201913 Jul 2019


Conference12th International Space Syntax Symposium, SSS 2019


Dive into the research topics of 'Investigating retail space performance through spatial configuration of consumer movement: A comparison of York and Leeds'. Together they form a unique fingerprint.

Cite this