Opportunity: Full-Time PhD In Partnership with dunnhumby

Full-Time PhD with an Industrial Sponsor
Company: dunnhumby
Location: London
Duration: 3 years Full-Time
Commencing: Sept/Oct 2020
Application deadline: 20 May 2020
Annual Stipend: £20,000 + tuition fees

We are pleased to offer a full-time PhD at the Bartlett Centre for Advanced Spatial Analysis (CASA), UCL, in partnership with dunnhumby, a world-leading customer data science company. dunnhumby analyse data and apply insights from shoppers across the globe to create personalised customer experiences in digital, mobile, and retail environments. They will provide access to billions of rows of data for this research. 

PhD Studentship Description
Gravity or Spatial Interaction Models have a long history in retail applications and for a long time they have been used to explore store revenues and sales in the context of their location. 

With the advent of store loyalty cards, retailers now know more about customers and their shopping habits over space and time than ever before. New channels such as online shopping and home delivery has meant that the ways that people shop are changing and the needs of retailers to understand these changes are more pressing. 

New theories and models of store and customer interaction need to be developed in order to maintain commercial viability. Is there still a place for traditional spatial interaction models or will new modelling paradigms need to be developed in order to be able to take advantage of detailed ‘big’ transactional data now routinely collected by retailers? 

Recent advances in computational artificial intelligence techniques coupled with visualisation have been developed in other sectors where spatial interaction models have had a traditional foothold, however to date similar advances have not been published for the retail sector offering potential for a rich seam of research to be exploited.

A nuanced understanding of store gravity will be developed, taking into consideration, amongst other things, the relationships that exist between store sales, the size and location of those stores and surrounding competitor stores, along with local demographic profiles and changes. Further research could test the robustness of such a model under restricted data scenarios. e.g. how reliable are predictions if we only have store information for specific competitors available? 

Desired outcomes could include: a parsimonious model with reduced store data able to predict as effectively as a fully calibrated model; a modelling and visualisation tool for evaluating the impact of store openings and closings.

Eligibility and how to apply: 
Applicants would normally be expected to hold (or nearing the completion of) an undergraduate degree in a relevant discipline (such as Geography, Planning, Economics, Political Science, Social Science, Computer Science, Mathematics, Statistics, Physics) at 2:1 or above. Masters degree or other relevant experience would be an advantage. This studentship is open to UK/EU and Overseas applicants. 

Please apply by sending a CV and separate covering letter outlining your motivation for wanting to pursue this PhD and initial ideas for the project (no longer than 1000 words) to Rhodri Jamieson-Ball by the closing date. Shortlisted candidates will be invited for interview in June. For any informal enquiries about the studentship, please contact Dr Adam Dennett at CASA.