A clustering study to verify four distinct monthly footfall signatures: a classification for UK retail centres 

Technical Report 1

Christine L Mumford*, Catherine R Parker**, Nikolaos Ntounis** and Ed Dargan**

*School of Computer Science & Informatics, Cardiff University

**Institute of Place Management, Manchester Metropolitan University

First published on January 27, 2017 

Overview

This report describes the application of K-Means clustering, Principal Components Analysis, and various statistical techniques to retail footfall data and verifies the existence of four distinct monthly footfall signature types, exactly as first proposed in the High Street UK2020 project. The footfall data was supplied by Springboard LTD and consists of hourly records broadcast from several hundred counters located in traditional retail centres throughout the UK. Before using the data, we checked its completeness for every counter by identifying any missing hourly data. The computed completeness figure at 96.38 % proved the counters to be very reliable. The identified retail centre types are: comparison, holiday, convenience/community and speciality. Comparison shopping centres tend to be located in the larger town and city centres and their monthly signatures can be identified by a footfall peak in December, coinciding with the Christmas preparation period. Holiday towns are busier in the summer months and footfall drops right down in the winter, whilst convenience/community centres tend to have more of a flat profile throughout all the months of the year. Finally, speciality centres seem to be somewhat of a “hybrid” type between comparison and holiday, insofar as they have peaks in the summer and in December, although these peaks are not as pronounced as they are in pure comparison and holiday centres.  

The article can be accessed in full using the top right hand link on this page.