Continuing on the theme of identifying research questions for our Improving the Customer Experience in Town Cenres: Bringing Big Data to Small Users project, we are wondering whether we can start to predict footfall in a particular location.
With more retail sales moving on-line and out-of-town then traditional catchment areas or numbers may need updating. In fact, in HSUK2020, Millington, Ntounis, Parker and Quin (2015) found that local resident population was a better predictor of footfall in smaller locations than catchment statistics. We like footfall as a measure as it concentrates on actual attractiveness (the number of people a retail centre actually attracts) rather than ‘potential’ attractiveness (catchment).
We will develop an improvement on existing methods of identifying catchment by providing a new method of predicting footfall (consisting, initially, of those components identified in HSUK2020, i.e., geographical location, location of nearest stronger centre, resident population, employment, tourism and vacancy rates).
First of all, we will want to compare estimated footfall from our HSUK2020 footfall predictor against actual footfall and catchment data for all towns in the Springboard historical footfall dataset. Then we will refine our original HSUK2020 model to improve its predicative ability. This will also allow decision-makers to estimate how attractive a location is (how many people it should be attracting) – especially important for smaller centres that may not be able to afford to collect real-time footfall data.
*This entry was first published on Prof Cathy Parker’s blog