Home » Archives for October 2018

2018 Masters Research Dissertation Programme Winners Announced

The winners of the CDRC’s 2018 Masters Research Dissertation Programme were recently announced at the Data Analysts User Group (DUG) Conference.

The CDRC-led programme provides the opportunity for students to work directly with an industrial partner and commercial datasets. This year, 20 students participated in the programme with the support of a range of partners, including Sainsbury’s, Shop Direct, Boots and M&S. Short summaries of the projects have been provided by the students and can be found here.

A panel of industry judges awarded three prizes based on the calibre and impact of the research. This year’s panel comprised Matthew Pratt (Javelin), Thomas Murphy (The Crown Estate) and Helen Parker (Tesco). The judges commented that the standard of work was very high and noted that a wide range of new techniques for handling big data were well implemented across the programme.

Prize Winner

Lenka Hasova (University of Liverpool) and Boots

Investigation of prescription flows between General Practices and Boots pharmacies via spatial interaction modelling techniques

Lenka’s study applied a range of different model specifications to estimate and analyse the distribution of GP-Boots pharmacy prescription flows in Merseyside, North-West England. The judges commented that the analysis was thorough and logically presented. Crucially it also tackled an important study area and her methods could be used to assist and improve the provision of pharmacies.

A short summary of Lenka’s research can be found here.

Runners up

Sophie de Kok (UCL) and Sainsbury’s

Predicting customer quality based on early shopping behaviour in online grocery retail

This project considered the exploration and evaluation of models for estimating the longer-term value of customers based on transaction data. The judges described Sophie’s submission as an advanced piece of analysis that was coherently presented. The research was well executed and clear benefit to the sponsor

A short summary of Sophie’s research can be found here.

Thomas Statham (University of Liverpool) and A National Broadband Provider

Forecasting network faults with Bayesian spatiotemporal statistical models

Thomas looking into means of forecasting network faults that might occur across the broadband infrastructure. This considered both time-series data provided by the sponsor and local sociodemographic factors. The judges commented that the dissertation demonstrated a thorough understanding of statistics and forecasting which were evidently well applied to tackle a complex issue in a novel way.

A short summary of Thomas’ research can be found here.

Poster Winner

Kahina Ait Ouazzou (UCL) and Movement Strategies

Analysing Customer Behaviour for Shopping Centres in London Using GPS Data

A ballot at the DUG conference was held to award the poster prize. Kahina’s project developed a conceptual framework on how to use GPS data to capture customer behaviour for shopping centres. It considered anonymised GPS data for Greater London in 2017.  Through the identification of shopping centre visits and home locations from GPS data, it was possible to analyse customer behaviour at a high spatio-temporal granularity and define the retail catchment areas.

A short summary of Kahina’s research can be found here.

 

Photo Credit: Jan Wright Photography

Inclusive and healthy mobility – new webpage launched

We have launched a new webpage for the ESRC-funded project ‘Inclusive and Healthy Mobility: Understanding Trends in Concessionary Travel in the West Midlands’. The project analysed 400,000 anonymised concessionary passengers to understand the sharp decline in bus boardings in the West Midlands Combined Authority between 2011 and 2016.

The webpage provides

  • findings and evidence on recent boarding trends;
  • an interactive dashboard of regional boarding trends;
  • data packs and interactive maps of mobility and accessibility indicators;
  • manuals to reproduce the linkage of smart card data following the UK ITSO format.

Key findings include

  • Bus boardings under the ENCTS scheme which entitles senior residents to free bus travel, have fallen by 21 per cent between 2011 and 2016;
  • Only parts of this trend can be attributed to recent eligibility restrictions to the scheme;
  • Our smartcard analysis identified six types of passengers with distinct residential patterns within the West Midlands Combined Authority;
  • A sharp withdrawal of regular bus users occurred in 2015-2016 and can be geographically linked to social disadvantage;
  • The decline in patronage coincides with structural changes in the transport system (extension of Midlands metro, introduction of e-hailing services) and uptake of online shopping among senior residents.

More information and detail findings can be found here.

CDRC co-hosts Data Analytics, Innovation and the Impact on Business and Society event at Senate House

As part of UCL’s Impact Series the CDRC in partnership with UCL Innovation & Enterprise and the UCL, Bloomsbury and East London Doctoral Training Partnership  organised a networking event at Senate House, University of London yesterday (10/10/18).

The event titled ‘Data analytics, innovation and the impact on business and society’ sought to bring together academics and businesses in order to discuss key trends in data analytics and means through which partnerships between the two sectors can be forged.

The afternoon event included an impressive line-up of 17 academics, doctoral students and key industry speakers. They included a keynote from CDRC’s Dr James Cheshire who discussed geospatial visualisation techniques and introduced the work of the CDRC. Drawing on the theme of better collaborations between academia and the private sector, three speakers from industry (Virgin Media, Local Data Company, Movement Strategies) and one from academia (Kim Cassidy, Edge Hill University) presented their views on the CDRC Masters Research Dissertation Programme, having all been external partners at some stage. All four speakers described previous research projects and commended the impact that the programme has had on their own work. Following this four PhD students from the CDRC and the Bartlett Centre for Advanced Spatial Analysis described their research with commercial sector data.

The day also included a fascinating panel discussion chaired by UCL visiting professor Matthew Hopkinson. The panelists included Peter Marks, CEO, Deltic Group; Sophie Birshan, Head of Retail Analytics & Insight, John Lewis & Partners; Graham Seaton, Property Asset Manager, Ann Summers and Alex Loizou, Co-Founder, Trouva.com. The debate covered issues and opportunities that arise from the changing nature of Big Data and how retail has had to adapt. Finally, the event was concluded by CDRC Director Paul Longley.

The event was attended by 55 contacts from industry and a large number of academics. It is hoped that the event will help encourage more external collaborations either in the form of short-term masters dissertations, or longer-term research projects and data licence agreements.

*Matthew Hopkinson, Visiting Lecturer at University College London and facilitator of the event, has written a very interesting post event blog on ‘Why Academia and Business can learn more from each other‘.

 

CDRC and AGI host the first Geo+Data London event

The Consumer Data Research Centre (CDRC) in partnership with the Association for Geographic Information (AGI) hosted the first Geo+Data London event at University College London (UCL) on 2 October 2018.

The seminar series is designed to bring together academia and industry in order to share and learn from experiences and best practices in the use of Big Data. The first event featured Alistair McMahon, Analytics Director at Telefónica UK/O2, and Balamurgan Soundararaj, a doctoral student in the Department of Geography at UCL.

Alistair McMahon’s talk, entitled ‘Smart Cities and Smart Transportation’, focused on the important connectivity work that O2 engage in, concluding that a data led approach to planning was critical to the success of smart cities.

Our academic speaker, Bala Soundararaj, gave an illuminating talk on the particular challenges he has encountered during his research ‘Estimating real-time high street footfall from Wi-Fi probe requests’, as well as explaining the novel problem-solving methods he deployed. A summary of our speakers’ presentations and their slides will shortly be added to the Geo+Data London’s dedicated webpage.

The next Geo+Data London seminar will be on 4 December 2018 and will feature James Ketteley from Mastercard and Justin van Dijk from UCL. Registrations for this event will open shortly.

Packed house at the first Geo+Data London event.