Home » Archives for April 2016

CDRC School Event: Profiling Neighbourhoods and Understanding Consumer Habits

We welcomed students from Hemsworth Arts and Community Academy to the Centre for a workshop on Consumer Data last week.  The session, which was voluntary, was attended by students from various disciplines including Geography, Psychology, Maths and Law.

The workshop started with a short lecture from CDRC’s Andy Newing on Big Data and the work of the Consumer Data Research Centre, followed by an exercise where students learnt about Geodemographic classifications and used CDRC maps to explore their local area.

Session 1

The students were split in to small groups and each allocated a Supergroup from the 2011 Census Output Area Classification (OAC) for the UK (Cosmopolitan, Ethnicity Central etc), using CDRC maps they identified where in West Yorkshire those groups resided.

 

Once they had identified the areas, the group used google street view to explore the area and make observations about the type of people that may live there.  They discussed their observations in groups:

  • If there were lots of security alarms or bars on windows, did this indicate a high crime rate in the area?
  • Did the number of cars indicate that most people drive to work?
  • What kind of jobs they thought the residents may have

The groups then returned to CDRC maps and compared their observations with data collected in the 2011 Census, using the following maps:

The students then used the information they had found to decide which consumer brands are most likely to be popular among these households and which retailers these consumers are most likely to shop at. The interactive session ended with a short discussion about how commercial organisations (such as retailers) or local government may use these forms of small area classification to support decision making.

Session 2

In the second session the groups participated in three short activities:

 

Activity 1 – Tour of CDRC

The students were taken on a tour of the CDRC and Leeds Institute for Data Analytics, as well as meeting researchers they particularly enjoyed exploring data visualisations on the Omniglobe.

 

Activity 2 – Temporal sales variations

Students worked with Research Postgraduate Tom Waddington to explore temporal sales patterns at the level of an individual retail store, considering the impact of area characteristics on these store trading characteristics. Students were able to explore simulated store-level data that were similar in ‘look’ and ‘feel’ to the type of consumer data that Tom uses for his research.

 

Activity 3 – Sales variations by geodemographic characteristics

Research Postgraduate and GIS Teaching Assistant Nick Hood worked with small groups of students to explore expenditure rates by geodemographic classification. Nick presented a range of topical data which highlighted variations in expenditure by OAC ‘Supergroup’ and product type and students contributed some informative suggestions to explain observed expenditure variations.

Do only affluent consumers buy green-labelled products?

By Professor William Young, Chris Brown (Asda) & Andy Peloe (Callcredit)

Combining sales and demographic data can provide better insight into consumer behaviour.

The current evidence on consumer buying behaviour and sustainability issues relies on attitudinal, self-reported or national sales data. This is often not close enough to real behaviour nor helpful at a company level.

With access to sales and demographic data, we are now seeing the detail at a geographical, product or consumer type level that can only help decision making. But this can only be done through exciting collaborations across sectors.

The Consumer Data Research Centre, Asda and Callcredit came together to explore if there are links between the affluence of consumers and green-labelled products they buy. Current wisdom says green-labelled products are bought by those who are richer. Of course, we have to keep in mind when looking at this topic that retailers already choice edit some product categories providing only products with green labels.

Customer views

Asda has already been asking their customers views on sustainability issues for a number of years through their Everyday Experts panel, which is the UK’s largest survey of its kind with 20,000 members. The company wanted to take this further and worked with CDRC investigating some product categories not affected by choice editing to explore consumer behaviour.

Callcredit has a partnership with YouGov and used its survey data on green issues to model and build a green and ethical segmentation product. Callcredit has worked with CDRC to explore how this could be used in combination with other datasets.

To explore the affordability question, four datasets were used:

  • product sales from Asda;
  • green and ethical segmentation data from Callcredit;
  • census data from the Office for National Statistics; and,
  • the index of multiple deprivation for England from the Department for Communities and Local Government.

The datasets have been linked to the supermarket location using the lower super output area (LSOA) in which each store is located. A LSOA is a small area geographical unit which contains approximately 1,000 individuals.

Age and gender impacts

Through regression techniques analysing the combined dataset, the results clearly indicate that for the same population the drivers and the barriers for the consumption of products vary by product. For example, for one product category those segmented as “rich” and “eco-friendly” generally purchase the green-labelled products. This is what we expected but we also found that demographics such as age and gender have positive influence.

For another product category, the results of the model indicated that purchase behaviour is influenced by green attitudes and store type but not demographics.

This is the first time a large sample comprising of actual sales data of products across a number of stores and linked to an equally large number of data points from other sources has been used. It provides valid and reliable insights into drivers and the barriers of the consumption of green-labelled products, making the results more authentic. Hence, the results clearly indicate that the consumption behaviour and the affluence of consumers varies between product categories and other factors are at play.

Sales predictions

The results of this work are still being analysed but initial conclusions have provided good outcomes for the companies. The model developed from the combination of these datasets can be used to predict the sales of products with green labels at existing and new stores using the demographics of the local population.

This will help to better target customers of particular stores with green-labelled products they are most likely to buy, rather than an all or nothing offering. CDRC was formed to conduct this type of research with companies to find exciting new insights on important societal issues, such as ethical and sustainable consumption.

Prof William Young is a Co-Investigator at the ESRC Consumer Data Research Centre.  He also leads the Sustainability Research Institute at the University of Leeds. 

Dr Chris Brown is senior director sustainable business at Asda Stores. He joined Asda as an agriculture development manager with a remit to develop Asda’s strategies and activities across all sectors of agriculture. His role was extended to be head of ethical and sustainable sourcing before he was appointed to his present position in 2013, covering waste/resource management, communications and sourcing.

Dr Andy Peloe is a concept manager at Callcredit. Prior to his current role, Andy was responsible for the development of all Callcredit’s geodemographic products both in the UK and 40 international markets. As a concept manager, Andy’s focus is to seek out new data and technologies to drive forward Callcredit’s data and innovation strategy.

CDRC host LondonR for R enthusiasts

On 5 April 2016 the Consumer Data Research Centre (CDRC) along with Mango Solutions, hosted ‘LondonR’, the UK’s premier meetup group for users of the R statistical programming language. This free event runs quarterly at the University College London, with a series of workshops and presentations followed by an evening of networking at the in-house bar.

The event kicked off with a workshop led by Gabor Csardi, Principal Consultant, Mango Solutions, who demonstrated the basics of R packaging, finding and browsing code of R packages on GitHub and CRAN as well as best practices for R users.

The evening session was at full capacity, with R enthusiasts in attendance for presentations by:

  • RosettaHUB | the next generation data science platform – Karim Chine, RosettaHub
  • Highlights from the Shiny Developer Conference | Aimee Gott, Mango Solutions
  • Professional matchmaking in R | Building a simple, practicable recommender with R for matching expert professionals to those seeking expertise – Duncan Stoddard, DS Analytics

The next LondonR is scheduled for June 2016; the exact date will be posted on our website soon. For all past and present LondonR presentations, click here.

The Consumer Data Research Centre also delivers courses on R – our next course ‘Intermediate and advanced R for spatial data’ will be hosted at the University of Leeds. For more on this and all other training by the CDRC, click here.

For more on Mango Solutions.

Gabor Csardi kicks off LondonR with a workshop
Gabor Csardi, Principal Consultant at Mango Solutions, kicks off LondonR at UCL with an afternoon workshop

Leeds Digital Festival 2016

The Leeds Digital Festival is a multi-venue, city-wide festival celebrating digital culture in all its forms, which will take place throughout the city from Monday 25th to Friday 29th April 2016.  The CDRC will be hosting, speaking and taking part in a number of events over the course of the week.

Monday 25 April

Sustainability Webinar

We’ll be starting the week with a joint event with some of our partners.  Speakers from Asda, Callcredit and Samsung will join Prof William Young for a webinar to discuss ‘How can Big Data help deliver sustainability strategies?’

During the one hour session, the panel will discuss how Big Data is currently utilised to help deliver successful sustainability initiatives within Asda, Callcredit and Samsung.  The panel will also participate in a live Q&A session with the audience.

Find out more


The City Talking: Tech in Leeds

On Monday evening we will be attending the premiere of ‘The City Talking: Tech in Leeds’ which features CDRC Director Mark Birkin.  The hour long documentary explores the heritage and current vibe of the Leeds tech scene.

 

Wednesday 27 April

On Wednesday a number of our researchers will be helping out with the Vision Zero Innovation Lab, which is being held by Student Data Labs at the Leeds Institute for Data Analytics.

The aim of the Vision Zero Lab is to explore ways to bring the number of fatalities on the road to zero.  Attendees will be learning practical data skills whilst tackling road traffic casualties. Come along and learn practical data science skills whilst working on projects that tackle real life problems.

Find out more

 

Thursday 28 April

CDRC’s William Young will host the second seminar in the LIDA series:  ‘Are we really green? Evidence using big data’

The seminar is open to all and will include an update on the CDRC Ethical and Sustainable Consumption Driver Project, a short presentation on encouraging individual to choose more environmentally friendly products and a feature presentation on Environmental Footprints: Consumption-based approaches to measuring environmental impacts.

Find out more

 

Further information

Visit the Leeds Digital Festival Website or follow @leedsdigifest16 for further information on the events and activities being held.

Is the environment conspiring against us to make us fat?

 

MM

Michelle Morris
Director of the ESRC Strategic Network for Obesity

 

Overweight and Obesity are a huge problem worldwide. The cost to the UK NHS is £5.1 billion annually with £11.5 billion annual cost to wider society. This does not account for the significantly reduced health related quality of life for overweight or obese individuals.  Overweight and obesity are also closely associated with type 2 diabetes, cardiovascular disease and certain cancers, which place further burden on individuals and society. It is a serious situation and one which is not going away quickly.

So what causes overweight and obesity? In the simplest terms they are a result of eating too much and not moving enough. However, in the real world of everyday life it is a much more complex and multifaceted relationship. Sometimes it feels like our environment conspires against any best laid plans to improve diet and increase physical activity, thus reducing or preventing incidence of overweight and obesity. Imagine these scenarios:

  1. You forget to take your lunch to work, the local café or corner shop often lack healthy options. Fish and chips are on offer again and you know that will keep you full all afternoon.
  2. You end up working late and miss the circuit training class you had planned to attend at the gym. You are so hungry and tired that you decide to pick up a take away from the Pizza shop you pass on your way home.
  3. You think that cycling to work would be nice and a good way to build activity into your day – but wait, it is often raining, the roads are really busy and you are not confident with crossing that big intersection where there is no bike lane. You could cycle on the canal tow path, but it is dark and there have been a number of attacks there and you don’t feel safe.
  4. Your office is on the 5th floor. That is 100 vertical steps. It would only take a few minutes each day and contribute to increased levels of physical activity. However, the lift is just there, I might have to wait, but I’m tired and I really can’t be bothered with the stairs.

I think most of us can relate to all of these scenarios. Ultimately as an individual we make the choices but often we may feel like there are a number of other people, situations or organisations at fault. This is why to better understand and ultimately tackle overweight and obesity many different types of organisations need to come together. In an era of growing volumes of data on all aspects of human behaviours it is timely to bring big data analytics into the forefront of obesity research.

It is exactly this that the ESRC Strategic Network for Obesity is trying to achieve. This international network brings together experts from many disciplines within and beyond academia to explore how we can use data about us and our environment to table overweight and obesity. How can we improve the environment so that it is easier to be physically active and eat well than it is to jump in the car and nip to a take-away? Hopefully we can find answers to this question through our network meetings and related activity.

The network will host four meetings over 12 months. The first meeting was in Leeds in November, with the second taking place in Cambridge last month.  If you missed the meetings you can catch up on activity by reading our storify or watching the seminars.    Follow us on Twitter: @obesity_network.

By Michelle Morris, University Academic Fellow. Twitter: @drshellm  @obesity_network.

 

This blog was originally posted on the University of Leeds – CSAP Blog.

 

How useful is Big Data in the fight against Obesity?

International experts came together in Cambridge just before Easter to explore how Big Data can be used to tackle obesity and obesity related diseases.  The audience of 60 consisted of academic and non-academic experts, with wide ranging interests including big data, nutrition, health and geography.

The second of four meetings, this meeting focused on Data, Methods and Models.  The morning session included presentations from network members followed by a panel discussion.

Presentations

Dr James Woodcock (CEDAR) & Dr Robin Lovelace (CDRC) discussed modelling and visualising large datasets to guide active policies. They presented a case study on the new Propensity to Cycle Tool, to demonstrate how novel forms of data and Big Data can serve public health through promotion of active travel.

Key points covered during the presentation included the issue that information needed for understanding travel behaviour and health are not at consistent geographies (Dr James Woodcock) and the need to join up public health/obesity research with training in coding, i.e. if you want to ‘do’ big data, you need to know how to code (Dr Robin Lovelace).

Dr Darren Greenwood from the University of Leeds covered interpreting results from analysis with Big Data and provided a number of interesting examples from epidemiology.  He covered the common pitfalls and encouraged the audience to think about what makes data ‘big’.

Prof Adam Drewnowski, Director of the Center for Public Health Nutrition at the School of Public Health provided an international perspective on Big Data and the Obesity Epidemic.  He explained why the commonly presented state level obesity maps are not helpful and presented some interesting findings on the links between property prices and obesity.

The presentations and videos are available here.

Panel Discussion

Dr Seraphim Alvanides, Dr Daniel Lewis and Dr Sandy Tubeuf joined the morning’s speakers for a lively panel discussion before lunch, exploring the practicalities of using Big Data in the fight against obesity.  Key points covered in the panel session include:

  • Issues around identification of subjects in data analysis with GPS coordinates discussed. Discussion included methods of analysis and the importance of masking location in code which is shared in addition to results which are disseminated. The importance of confidentiality and consent where applicable was highlighted.
  • There is chance of generating erroneous inferences from ‘big data’ if robust analysis methods are not used. Discussion focussed around whether it is better to be ‘roughly right’ rather than ‘precisely wrong’.
  • ‘Big Data’ invites opportunity for ‘hypothesis free’ research. Discussion followed around the need for a priori hypotheses and whether ‘hypothesis free’ analysis is good research practice or just a ‘fishing expedition’.

Attendees took advantage of the winter sun and took a short walk around Cambridge, before ending the day with a networking lunch.

Dr Pablo Monsivais, who chaired the meeting commented on the success of the day: ‘I was pleased to see such a high level of interest and engagement from the audience, over half of which were not network members.  We need fresh, joined-up approaches to tackling obesity and the talks and workshop illustrated how this strategic network is enabling important intellectual and methodological advances using big data.”

If you weren’t able to join us, you can follow the day via the Meeting Storify Board.

The next meeting will take place in London on Wednesday 18 May 2016.  The meeting will follow a similar format but will focus on Novel Results & Visualisation. Book online.