Home » News

Understanding Household Mobility – Nik Lomax

Why do we need to understand household mobility?

People move home for all kinds of reasons: to take up a new job, to up-size when they have a child or down-size when they retire, and after some kind of family event like marriage or divorce. Understanding these patterns of household mobility is extremely important for resource allocation and policy decisions: households need roads, parks and bin collections, children need school places, and everyone needs health service provision.

Mobility patterns also impact upon less tangible but equally important elements of our society: having strong community cohesion and a sense of place are important for everyone. When households move, they change the size and composition of the population in local areas. For example, a retired couple selling their house to a family with two children creates demand for additional school places but also changes the demographic composition of the area, severing established community bonds and forging new ones.

Internal migration is the process which delivers the most substantial change in population at a local level, results from the 2011 Census of Population reveal that ten times more people moved within than immigrated from outside the UK: 6.8 million people made an internal move in the year to March 27, 2011. Internal migration patterns are a crucial component in sub-national population projections which inform forward looking policy. We need timely data which provides good coverage to understand mobility patterns.

Isn’t there a register of migration?

There are several sources of data which report migration within the UK, but no single register like, for example, the one in place in Sweden. The most comprehensive (with the largest coverage) is the Census of Population, which askes for information on where every individual lived twelve months ago. These data are extremely valuable, but have limitations:

  • they are only available once every ten years;
  • data are not released for a substantial period of time after the census is completed;
  • data are released at pre-defined administrative boundaries (the smallest being Output Area);
  • they only capture transitions for a single year – the question asks ‘where were you living one year ago’ – so multiple movements that people make during the year are not recorded.

There is debate around the future of the conventional census, given the cost of administering the survey (estimated at £480m[1] for 2011) and the current political agenda for austerity.

Official estimates of migration, released annually, are produced using administrative data from the National Health Service. When an individual informs their GP that they have changed their address the migration record is updated accordingly. Again, this is a hugely valuable resource but problems with these data include:

  • under reporting, especially amongst certain groups (e.g. young men who don’t regularly visit their GP);
  • geography, as the most detailed spatial units available are local authorities;
  • timeliness as the data are only released once per year;
  • the data only report one year transitions.

Other administrative data exist, for example the Higher Education Statistics Agency data used to improve estimates of student migration in the official statistics. These administrative data cover sub-groups of the population, and are not inherently intended for the measurement of migration.

The third option is the use of commercial or Big Data resources. Some work has been undertaken previously on these emerging forms of data and their utility for migration research. For example, work by Dr Michael Thomas and Professor John Stillwell uses a commercial survey called Acxiom to assess patterns of movement and distances travelled for the population sample. But, to date, work on demonstrating the utility of commercial data as a viable alternative to census or admin data, is fairly limited.

This is why the opportunity to use a large scale commercial dataset, as presented in this series, is so exciting. Commercial data has the potential to enrich our understanding, but we are not yet at a stage where it can be relied upon as the only source of information.

 

Accessing property and migration data

The Consumer Data Research Centre has partnered with online property search provider Zoopla and data insight consultancy Whenfresh to obtain data about the characteristics of properties which have been sold in England and Wales. For the 2014 calendar year, there were over 900k unique property transactions.

This data is available to researchers via the CDRC Secure Service:

View WhenFresh/Zoopla Property Rentals and Associated Migration Metadata

View Property Transactions and Associated Migration Metadata

[1] http://www.ons.gov.uk/ons/guide-method/census/2011/census-data/faq/2011-census–frequently-asked-questions.pdf

About the author

Dr Nik Lomax is a University Academic Fellow at the University of Leeds, his research focuses on the way in which demographic behaviour changes over time and how people interact with the areas in which they live and work. Much of his work focuses on the dynamic processes involved in migration but he is also interested in the social implications of changing demographic composition: household formation, social exclusion and population ageing for example. Areas are shaped by changing economic conditions, policy interventions and social attitudes, which in turn has an impact on demographic behaviour. Modelling and explaining these complex interaction is key to the work which he does.

Dr James Cheshire wins the prestigious Royal Geographical Society Cuthbert Peek Award!

The Royal Geographical Society’s medals and awards “recognise extraordinary achievement in geographical research, fieldwork and teaching, photography and public engagement.” CDRC Deputy Director James Cheshire was among 17 recipients this year, winning the Cuthbert Peek Award ‘for advancing geographical knowledge through the use of mappable Big Data’.

Congratulations James!

CDRC Data Partner Forum 2017 – Presentations Available

On Thursday 25 May 2017 the Consumer Data Research Centre hosted its annual Data Partner Forum at the Leeds Institute for Data Analytics.

The aim of the event was to bring together existing and potential data partners along with interested academics to discuss current and future research opportunities. Led by CDRC Directors, Professor Mark Birkin and Professor Paul Longley, the event had an excellent mix of attendees from both consumer related organisations and academia.

The programme included presentation sessions on ‘Business Analytics’ and ‘Cities’, as well as a poster session, breakout discussions and a concluding panel session.  The main sessions included short presentations by a number of our researchers discussing the ways in which they are utilising consumer related data to better understand and explain patterns of social behaviour.  A number of the presentations are now available to view and download:

Energy

The geography of domestic energy consumption – Anastasia Ushakova (UCL) and Ellen Talbot (University of Liverpool)

Transport

The propensity to cycle tool: from conception in the clouds to implementation on the ground – Robin Lovelace (University of Leeds)

Using Travel Smartcards to understand trends in concessionary travel in the West Midlands – Jens Kandt (UCL)

Mobility

Quantifying the ambient population – Nick Malleson (University of Leeds)

Smart Street Sensor Project – Roberto Murcio (UCL)

Health

The role of big data in health research: a case study of obesity – Michelle Morris (University of Leeds)

Should you be interested in attending the event next year, or wish to discuss ways in which your organisation can engage with the CDRC, please contact Paul Evans.

 

What is research impact and why is it so important?

Rachel Oldroyd, UK Data Service Data Impact Fellow and Quantitative Human Geographer at the University of Leeds, shares her perspective on research impact.

Impact – a word that is used so frequently in academic and non-academic circles alike, but what is it and why is it so important? With so many definitions, the meaning varies greatly across institutions and disciplines, but there seems to be a certain agreement around one broad definition. Impact is the effect that research has beyond academia. Whether that’s how the understanding of a complex social theory affects policy and improves community life, or how a newly developed vaccine reduces the spread of disease among a population. Impact is certainly not a new concept, however its incorporation into the 2020 Research Excellence Framework (REF) has initiated a shift away from pure research excellence towards a focus on research dissemination and the effect on wider society.

Certain channels of dissemination are frequently used by academics; presenting at national and international conferences, publishing in academic journals and delivering research group seminars. Early career researchers soon realise that these activities are not really optional, participation is encouraged and expected during the life course of a PhD and beyond. But these activities are often not enough to create tangible impact. They have limited reach, restricted to academic audiences. So what can early career researchers do to increase the impact of their research and engage non-academic audiences?

I’m certainly no expert on the topic, I started a part-time PhD just over 18 months ago having spent three years as a teaching fellow at the University of Leeds. Impact was a term I heard frequently in research discussions and meetings, but as I was based in a teaching focussed role, it wasn’t something I had given much thought. However, as I started my research around foodborne illness, I realised the project had potential to make a difference and I began to think about ways in which I could develop impact. I started to identify people that would be affected and the ways in which the findings would be important. I believe identifying key stakeholders at an early stage is an important step in developing an impact journey. Key parties can not only help to disseminate the research beyond academic audiences but they can also provide valuable feedback regarding the research itself.

Stakeholder feedback can help to effectively outline and understand the problem, guide and mould the methodologies and also aid interpretation of results and policy considerations. Involving key parties throughout research development ensures the research remains timely and useful, but stakeholder engagement can often be difficult to establish. Taking advantage of existing contacts and using professional events as opportunities to network are ways in which early career researchers can reach out to key parties. I was particularly interested in talking to my local food safety team, so I sent a speculative email outlining my methodologies and they were happy to meet to start an open ended conversation about the research.

Although starting open-ended research conversations can only be a good thing, stakeholder engagement can bring about new challenges for researchers. Mainly, how to communicate technical theories to public and professional audiences? Explaining complex methodologies to non-academic audiences in an accessible way is often difficult. Although most Universities provide training to support early career researchers engaging stakeholders and the public, it takes much practice. Avoiding complex language and acronyms and focussing on results rather than methodologies are ways in which to make the research more understandable for non-academic audiences. Often public and professional audiences are less concerned with the minutia of the methods and more interested in results, outcomes and policy implications.

Alongside engaging with key stakeholders, developing research awareness is a crucial step in creating research impact. Writing a research blog, presenting at public seminar series, engaging with social media, and running public engagement events such as schools outreach sessions, are excellent ways to increase the visibility and profile of the research. However, life as an early career researcher is busy, and maintaining a balance between activities which promote the research and undertaking the research itself is one of the main challenges in developing impact. Impactful activities certainly shouldn’t detract from research excellence and learning to say no is also a vital skill that early career researchers should learn.

It is important that impact development is considered at all stages of the research so it remains timely, useful and policy relevant. Although the concept of impact isn’t new, as the 2020 REF looms, the emerging trend away from pure research excellence towards research dissemination and measured societal effect will not only demand established academics to change their approach to research, but it will also require an increased level of support for early career researchers entering this new landscape of academia.

Find out more about Rachel’s research – Using novel types of data to detect illness caused by contaminated food or drink.

Oxford Retail Futures Conference – call for papers

The changing role of place in a digital world

Background

The continued importance of physical locations for retailing has come into question, as the ways in which shoppers choose to shop are changing in fundamental ways. Three important drivers of this change are urbanization and the attendant growth of an increasingly demanding and affluent middle class (especially in emerging markets), the impact of technology, and the effects of regulation. As a consequence, pressures are beginning to affect the traditional roles of stores and the requirements for particular formats of stores are being revisited.

This year’s conference considers the implications of these changes for the physical landscape of retailing. We will address in particular how retailers’ space requirements are likely to change; to what extent the role of physical stores needs to be reimagined by retailers; how this will affect business models, local economies and communities; and how these changing roles will vary by sector as well as by geography; and what is at risk for stakeholders in retail property markets as well as for policymakers at national and local level.

Topic Selection

In this call for papers or extended abstracts (minimum 1 page of A4), we would like to invite contributions in areas related to the changing physical attributes of the retail sector. These may include theoretical and conceptual work, as well as examples from practice. Empirical results of analysis of large data sets are also welcome.

The call is focused, non-exclusively, on the following topics (applied in the retail context, both at the store-end and in the extended retail value/supply chain).

• Advances in the sourcing and analysis of spatial data
• Analysis of the impact of retail distribution activity
• Case studies of the evolution of retail places
• Geographical aspects of customer shopping behaviour
• Innovation in retail business models
• Local and regional economic studies and impact analysis
• New methods and tools for spatial analysis
• Policy implications of the spatial evolution of retailing
• Retail planning and development strategies
• Strategies and policies of retail real estate investors and developers
• Supply chain consequences of emerging retail distribution networks
• The evolution of the planned shopping centre
• The geographical consequences of omnichannel retailing
• The measurement of local economic activity
• Town centre development

Papers submitted will be reviewed by the academic board. Extended abstracts and work in progress are welcome.

Deadlines

• submission of extended abstracts – 31st July 2017
• notification of paper acceptance – 31st August 2017
• submission of final papers/extended abstracts  – 1st December 2017

Members of the Conference Academic Board

• Dr Richard Cuthbertson, OXIRM, Saïd Business School, University of Oxford, UK
• Dr Jonathan Reynolds, OXIRM, Saïd Business School, University of Oxford, UK

Contact Details

The conference is being organised jointly by the Oxford Institute of Retail Management, Saïd Business School, University of Oxford and the Consumer Data Research Centre (CDRC).

Further details can be found here.

For academic-related enquiries, please contact Dr Richard Cuthbertson at richard.cuthbertson@sbs.ox.ac.uk or Dr Jonathan Reynolds at jonathan.reynolds@sbs.ox.ac.uk

For conference logistics, please contact Claudia Seiler at claudia.seiler@sbs.ox.ac.uk, OXIRMEnquiries@sbs.ox.ac.uk

Please send papers / abstracts to OXIRMEnquiries@sbs.ox.ac.uk

Registration fee

£195

The fee can be waived for students and presenters. Therefore, please use the dedicated students or presenters registration link.

All other delegates please use this registration link.

CDRC paper wins CASA prize for the Best Paper on Spatial Analysis at GISRUK 2017

A CDRC funded conference paper by Luis Francisco Mejia Garcia (former CDRC Masters Research Dissertation student, now BlueOptima), Guy Lansley (CDRC, UCL) and Ben Calnan (Movement Strategies) won the UCL CASA prize for the Best Paper on Spatial Analysis, in memory of Sinesio Alves Junior, at this year’s GISRUK conference. The prize was judged by an independent panel of academics from the Centre for Advanced Spatial Analysis (CASA). The paper entitled “Modelling Spatial Behaviour in Music Festivals Using Mobile Generated Data and Machine Learning” aimed to demonstrate the potential utility of mobile phone data to understand and predict spatial consumer behaviour within an enclosed environment. The study was an extension of research conducted for a dissertation in last year’s CDRC Masters Research Dissertation Programme.

The extended abstract is available on the GISRUK website: http://manchester.gisruk.org/proceedings.php and Luis’s Masters Research Dissertation case study here.

CDRC paper wins prize for Best Paper by a Young Career Researcher at GISRUK 2017

A CDRC conference paper by Alyson Lloyd and James Cheshire won the GISRUK 2017 prize for best paper by a young career researcher, sponsored by Pixalytics. The presentation entitled “Challenges of Big Data for Social Science: Addressing Uncertainty in Loyalty Card Data” was voted for by attendees of the conference. The research aims to understand the prospects of loyalty card data for uses within social science using the CDRC’s High Street Retailer data. The paper presented the construction of a method to detect unreported customer address changes in the data, by analysing their store visiting behaviours.

The extended abstract is available on the GISRUK website: http://manchester.gisruk.org/proceedings.php and a case study on Alyson’s research can be found here.

CDRC Maps London’s Daily Commute

Oliver O’Brien, CDRC Data Scientist and Senior Research Associate at UCL Dept of Geography, has published a visualisation of Department for Transport traffic counts by mode, for the morning rush hour for London, at http://vis.oobrien.com/trafficcounts/. Viewers can access raw counts data over the last 15 years, for several thousand measuring sites in the capital, and see how volumes and mode share have changed, on a street-by-street basis, as the capital’s commuting patterns continue to evolve.  A detailed description of the interactive map is at http://oobrien.com/2017/03/evolution-of-londons-rush-hour-traffic-mix/. The CDRC is looking to expand its coverage and work of transport-related consumer datasets, and is working with local councils to help analyse and visualise their transport-related datasets. The work was featured in a recent article published by Wired: http://www.wired.co.uk/article/london-traffic-map-busy-congestion has has already led to expressions of interest from other potential data providers.

The online tool that can track, monitor and analyse nutritional intake

Janet Cade, University of Leeds

We all know that most people could improve the quality of their diet. Most of us do not eat the recommended five-a-day portions of fruit and vegetables – let alone seven or even ten, as some have suggested. Nor do we consume adequate amounts of oily fish. Instead, intakes are often too high in saturated fats and sugars added to foods and fruit juice, and too low in fibre and some key vitamins and minerals, including vitamin A and iron. A significant proportion of adults in both the UK and the US are obese or overweight. Intake of red and processed meat is too high, and meat consumption continues to rise in the US, the European Union and the developed world. Despite a shift toward higher poultry consumption, the largest proportion of meat consumed in the US is still red meat (58%).

There are serious implications for long-term health as a result of this disordered way of eating. To improve the situation we need to know how much energy and nutrients are being provided by our food. To help do this, we developed myfood24, an online dietary assessment tool that can support accurate, detailed recording of food and nutrient intake by researchers, but which can also support patients with diet-related conditions, sports enthusiasts, families with “picky” eaters and others. With data on 40,000 nutrients, it includes the largest and most complete food composition table in the UK, and possibly the world.

Monitoring intake

The size of portions and packaging has increased over the past 50 years, as has the number of products on supermarket shelves. This variety of choice makes it hard for consumers to even start to estimate how many calories or nutrients they might be consuming.

A new generation of smartphone apps offer users a chance to monitor their intake. However, there isn’t strong evidence that most of these are effective. Twenty-eight of the top 200 rated health and fitness category apps from Google Play and iTunes focused on both weight management and self-monitoring diets. When these apps were compared to people using a standard record of weighed food that they ate, the apps over- or underestimated energy intake by 10-14%.

But it’s not just consumers who are affected by inaccurate monitoring. Researchers, who base their studies on this kind of data, also encounter problems.

A major limitation of nutrition research is getting an objective measure of dietary intake. Misreporting is a big problem when people self-report their diet and is particularly common in overweight or obese people. Misreporting generally tends towards under-reporting of unhealthy foods and over-reporting of fruits and vegetables.

Metabolic profiling, which involves testing urine for the hundreds of metabolites that provide chemical signatures of food and nutrient intakes, doesn’t require self-reporting and may be a useful addition to self-reporting. A highly controlled study of 19 people fed four different diets found differences in metabolite concentrations. While this approach cannot replace the need to determine what actual food and nutrients have been consumed, it could be used as an objective measure to validate self-reports.

Understanding what nutrients are in the food we eat also relies on having comprehensive and up-to-date food composition tables – standardised national databases with accurate measures of many nutrients in typical foods. Standard food composition tables in the UK list around 3,000 food items, the majority of which are generic rather than branded (which more of us are likely to consume). While they include the full range of nutrients, they only include a limited selection of foods which are available for purchase.

Useful data.
Shutterstock

Pre-packed foods legally have labels with nutritional values. These include values for energy (kJ and kcal), and amounts of fat, saturates, carbohydrate, sugars, protein and salt. Further information can be included but is not compulsory for mono and polyunsaturated fats, starch, fibre, vitamins or minerals. If a nutrition or health claim is made on the packaging then the amount of that nutrient must also be stated.

Real time feedback

Developed with funding from the Medical Research Council, myfood24 combines the convenience of new technologies with an enhanced food composition table. Covering a wide range of generic and branded foods, it’s a quick and easy tool to help researchers, and potentially also clinicians, to track, monitor and analyse nutritional intake. We mapped the 40,000 nutrients from food label information and generic food data. To get an idea of the scale of this, the number of products on supermarket shelves is around 50,000 items.

The tool replaces the need for time consuming and costly coding of paper records that researchers and clinicians use. It means that people can record their dietary intake by selecting foods and portion sizes and adding them to their food diary. We hope this will support more accurate self-reporting, especially as users can be less self-conscious than when reporting to an interviewer. Researchers can then use results from this to find out in detail what foods and nutrients are being eaten. This data can then be linked to health outcomes or matched against recommendations.

Real time feedback of nutrients in foods could help us choose a more appropriately balanced diet over the week. Much as we have come to rely on regular visits to the dentist to ensure our teeth are healthy, the regular use of dietary monitoring could help us to ensure that our food and nutrient intakes are also healthy.

Janet Cade, Professor of Nutritional Epidemiology and Public Health, University of Leeds

This article was originally published on The Conversation. Read the original article.

Hiscox University Data Challenge

Lawrence Ning Lu, CDRC PhD Student, is leading a team from the University of Leeds in the Hiscox University Data Challenge.

The ‘Hiscox University Data Challenge’ sees three teams from LSE, York and Leeds Universities compete to solve real world problems, giving Hiscox the opportunity to tap into creative and analytical minds with a different viewpoint to those currently working in the industry. For the students, they get to experience real world problem solving, access to industry experts, networking opportunities with current graduates on the scheme and potential Hiscox sponsorship.

The Challenges

The team have completed the first challenge, which saw them modelling the causes of railroad accidents in America and identifying factors that may increase liability.

They have now moved on to the second challenge which asks them to consider  what factors of success would look like for a start-up company.  For instance, ‘is there a ‘Silicon Valley effect?’  Does the age of the CEO influence a start-up’s success and how?

The Experience

The team, which includes four students from our MSc Consumer Analytics and Marketing Strategy, recently visited Hiscox offices in York, to present the results of their first challenge to a senior team and also had the opportunity to network with the underwriters and analysts.

It’s not all hard work though, there was time for team bonding at Five Guys too.

Good luck to the team, we look forward to seeing their progress in the competition.

To find out more about Leeds Data Science Society, visit their website or follow them on Twitter.