Date(s) - 21/05/2019
10:30 am - 6:30 pm
Categories No Categories
The ESRC’s Consumer Data Research Centre is delighted to invite you to its fourth annual Data Partner Forum. This year we will celebrate the Centre’s achievements over the last five years and spotlight our significant successes. As the CDRC moves into the next phase of funding, we will continue to generate pioneering research alongside our data partners and to influence policy and practice.
11:00 – 12:30 Session 1: Smart cities
- Professor Mark Birkin – Welcome
- Dr Yi-Chun Ou, University of Leeds – An understanding of interactions with smart objects and their consequences
- Dr Emmanouil Tranos, University of Birmingham – Modelling clusters from the ground up: a web data approach
- Professor Alison Heppenstall, University of Leeds – Uncovering hidden structures and patterns in cities through mobility data
- Lightning talks: 1/ Dr Kevin Minors, University of Leeds – What is a particle filter? How can smart cities use them? 2/ Natalie Rose, University of Liverpool – Exploring the impact of weather on retail sales and consumer behaviours using machine learning 3/ Dr Alex Coleman, University of Leeds – Extracting actionable insights from free text police data 4/ Professor Roy Ruddle, University of Leeds – Using visualization to profile data
- Questions and discussion
12:35 – 13:35 Lunch
13:35 – 15:10 Session 2: Mobility, migration and housing
- Dr Kate Pangbourne, University of Leeds – Shifting travel behaviour through personalised messaging: data, argumentation, attitude and personality
- Professor Alex Singleton, University of Liverpool – Contemporary geodemographic analysis and urban analytics
- Dr Minh Le Kieu, University of Leeds – Data-driven cities: bringing together machine learning and big data
- Lightning talks: 1/ Jeroen Bastiaanssen, University of Leeds – Job accessibility and employment outcomes: the role of public transport in the UK 2/ Jack Lewis, University of Leeds – Assessing the presence of ‘e-food deserts’ in the UK groceries e-commerce sector. 3/ Stelios Theophanous, University of Leeds – Synergy PRIME – Multi-level Modelling, Simulation and Visualization
- Questions and discussion
15:10 – 15:25 Tea and coffee break
15:25 – 17:00 Session 3: Health and lifestyle
- Dr Carmen Piernas-Sanchez, University of Oxford – Behavioural interventions to improve the quality of grocery shopping: working with retailers and clinicians
- Dr Emma Wilkins, University of Leeds – Assessing diet and physical activity in a university student population: A longitudinal novel data approach
- Dr Will James, University of Leeds – Local area estimation profiles and consumer attitudes
- Lightning talks: 1/ Victoria Jenneson, University of Leeds – Supermarket purchase data in population dietary research 2/ Francesca Pontin, University of Leeds – Physical inactivity; determining the demographic and geographic factors using smartphone data 3/ Dr Michelle Morris, University of Leeds – What do you think about researchers using your lifestyle information: the LifeInfo Survey
- Questions and discussion
17:00 – 18:30 Drinks reception
Dr Yi-Chun Ou – An understanding of interactions with smart objects and their consequences
Smart objects connect to the internet with a level of intelligence, such as smart wearables, Amazon Alexa, and Google Home. While smart cities promise a better and livable future, smart objects are some of the devices making this happen. It is therefore important to understand how individuals interact with smart objects and whether the interactions improve subjective well-being (i.e., hedonic happiness [experiencing pleasure] and eudaimonic happiness [achieving meaningful goals]) over time.
We aim to examine the interactions in a customer-experience perspective and explore different types of perceived value from the interactions, which are expected to form relationship quality and improve subjective well-being.
Dr Emmanouil Tranos – Modelling clusters from the ground up: a web data approach
This talk proposes a new methodological framework to identify economic clusters over space and time, which builds upon recent developments in data science. Specifically, we employ a unique open source of commercial, geolocated and archived webpages. We interrogate these data using data science techniques to build bottom-up classifications of economic activities using the textual data included in the webpages. We take a fresh look at an iconic London neighbourhood – Shoreditch – that is rich in technology and creative industries, and has become a leading digital and creative cluster over the past two decades. On top of contributing to the discussion around the SIC limitations, our work also addresses some of the spatial and temporal limitations of the clustering literature. Instead of adopting an administrative unit as the spatial scale of the analysis, the novel micro-data we employ enables us to capture a more realistic footprint of the Shoreditch economic cluster. Moreover, our data sources enable us to go back to 2000 and observe how the cluster has evolved over time. In short, this adds to the limited empirical literature (Ter Wal and Boschma 2011; Balland, Boschma, and Frenken 2015; Delgado, Porter, and Stern 2015), which moves beyond a pre-determined understanding of economic clusters in spatial, temporal and technological terms (Catini et al. 2015).
Professor Alison Heppenstall – Uncovering hidden structures and patterns in cities through mobility data
Understanding how individuals use cities remains a considerable challenge. The appearance of new rich data sets (e.g. mobility, sensor and social media) has allowed researchers to begin to answer questions about who, why and where individuals are in cities. However, extracting insights from these large data sets often requires the use of methods drawn from machine learning and network science. Using mobility data (taxi and bike share), this presentation will briefly highlight ongoing work within CDRC that aims to uncover hidden structures and patterns in cities.
MOBILITY, MIGRATION AND HOUSING
Dr Kate Pangbourne – Shifting travel behaviour through personalised messaging: data, argumentation, attitude and personality
An overview of ADAPT, to develop more effective ways of using persuasive messaging in sustainable mobility apps or other persuasive uses of mobile technologies.
ADAPT is now at the mid-point, having developed a dataset of existing Sustainable Travel Communications (the STCD) and analysed the content to quantify the argumentation types and values typically present in existing material. The next step, using message persuasiveness experiments to understand what factors predict the perceived persuasiveness of messages promoting walking and cycling is progressing well. What are the implications for app designers and travel planners? We are currently moving on to messages about bus use and about air pollution and health and will finish with a short discussion of the remaining research design challenges such as incorporating contextualising sensor data, such as air quality information, into messages.
An overview of the work of a 5-year programme of research funded by the EPSRC Living with Environmental Change Challenge Fellowship programme www.adapt.leeds.ac.uk/
Professor Alex Singleton – Contemporary geodemographic analysis and urban analytics
The availability of open geodemographic classification have been transformative in terms of scientific reproducibility, and have extended their use widely. For public sector applications, such methodological orientation has provide a more defensible position when operationalised within the context of public service delivery. However, such classifications are reliant on Open Data which are increasingly under threat from a variety of externalities that challenge availability in perpetuity. As such, this talk illustrates new hybrid approaches to geodmeographics that retain reproducibility, however, consider inputs from a wider data economy. This work is illustrated in the context of the new CDRC Internet User Classification and applications to retail analysis.
Dr Minh Le Kieu – Data-driven cities: bringing together machine learning and big data
Data science and data-driven modelling can provide insights into the movements of people in cities. This enables urban planners to access the dynamism of a more agile and responsive decision making practice considering the feedback loop from ‘big-data’ to reshape their future decisions. This presentation describes the methods and several specific cases where data science and data-driven modelling can support urban planners to make faster and more accurate decisions. It focuses on the use of big data, such as public transport Smart Card data and social network data, in machine learning techniques and city simulation to model people and vehicle movements in cities.
HEALTH AND LIFESTYLE
Dr Carmen Piernas-Sanchez – Behavioural interventions to improve the quality of grocery shopping: working with retailers and clinicians
Diet is an important determinant of health, and food purchasing is a key antecedent to consumption hence improving the nutritional quality of food purchases presents a clear opportunity to intervene. Findings from our recent systematic review of interventions implemented in grocery stores suggest that price manipulations, healthier swap suggestions, and perhaps manipulations to item availability change food purchasing and could play a role in public health strategies to improve health. However, the evidence base for interventions in grocery stores is still very limited. We are conducting a series of studies to examine the effectiveness of interventions based around healthier swaps on the quality of the food purchased and eaten as well as the short term effects on relevant health outcomes. This presentation will describe two studies to test interventions to reduce saturated fat (SFA).
Dr Emma Wilkins – Assessing diet and physical activity in a university student population: A longitudinal novel data approach
Starting university is an important time with respect to lifestyle and diet changes. Studies have shown this to be a time when students are likely to gain weight. Understanding the dietary and physical activity behaviours of new university students is important to identify factors that might contribute to weight gain.
At the University of Leeds, new first-year students in campus halls of residence are provided with a swipe card which (i) is pre-loaded with credit to cover two meals per day during the week, and brunch at the weekend and (ii) provides access to University sports facilities. Swipe card data spanning the first term of University (September 2016 – December 2016) were obtained for 920 students aged 18-24, including information on individual food purchase transactions (food name, price, time) and sports facility usage (facility used, time). Food purchasing and sports facility usage patterns were assessed at the level of individual students using data-driven techniques, including cluster analysis.
This study demonstrates a novel approach to assessing diet and lifestyle behaviours that overcome limitations of traditional approaches, which suffer bias from self-reporting, self-selection and drop-out, and often only assess diet over short periods or a small number of dietary components.
Dr Will James – Local area estimation of expenditure profiles and consumer attitudes
In common with the rest of the world, the UK is experiencing major shifts in dietary patterns, as evidenced by changing patterns of food and drink expenditure. Consumer attitudes play a major role in product choices, with views on environmental sustainability, health and ethics all influencing the expenditure profile of an individual. Understanding the local scale patterns of expenditure and consumer attitudes is therefore critical for assessing market stability and for forecasting future trends. Despite its importance, there has been relatively little research into the spatial patterns of expenditure and associated consumer attitudes in the UK.
In this research, we present new GIS datasets of household expenditure and consumer attitudes across Great Britain. All datasets are at the Local Authority District level (n = 380) allowing for spatial variation to be assessed. The technique of spatial microsimulation was used to estimate household expenditure of 103 categories of food and drink for each local authority district of Great Britain for the years 2008 – 2016. Additional datasets relating to consumer attitudes (supermarket shopping behaviours, attitudes to sustainability, environmental issues and food) were also generated using microsimulation, allowing the relationships between consumer attitudes and expenditure to be assessed.
Dr Kevin Minors, University of Leeds – What is a particle filter? How can smart cities use them?
In this talk, we present an application of a particle filter to a crowd simulation model in order to demonstrate data assimilation on an agent-based model. Generally, agent-based models run independently from reality once initial conditions have been set. This is a major hurdle for smart city models that are constantly receiving new information. Using a particle filter to assimilate the data in real time is one way to overcome this challenge. We demonstrate how this can be done using a particle filter receiving location information from a corridor-like crowd simulation model.
Natalie Rose, University of Liverpool – Exploring the impact of weather on retail sales and consumer behaviours using machine learning
Retail is one of the most important economic sectors in the UK and due to the rise of e-commerce and convenience culture has undergone significant changes in recent years. An understanding of the key influences on product sales can help retailers stay competitive in the current landscape. Weather is a key influential factor on both product demand and purchasing decisions, however, due to the complexity of consumer behaviours, the impact that weather has on sales can be particularly difficult to monitor.
Dr Alex Coleman, University of Leeds – Extracting actionable insights from free text police data
When a crime occurs, large volumes of information relating to the event are being recorded via free text police input systems. These data contain useful information that if appropriately collated, analysed, and understood could lead to significant reductions in harm. Using natural language processing techniques this project aims to demonstrate that actionable insights (e.g. identifying unobserved trends in incident type, context, or geographical locality) can be derived from police free text data. Using topic modelling and vector space model approaches it has been possible to identify a greater level of specificity in crime events when compared to traditional crime categories.
Professor Roy Ruddle, University of Leeds – Using visualisation to profile data
Analysts and researchers face major hurdles understanding the characteristics and quality of their data. This Alan Turing Institute (ATI) project is investigating how we can exploit data visualisation techniques to overcome those hurdles. If you would like to identify quick wins, share best practice or take part in our visual data profiling workshops then please contact the speaker.
Jeroen Bastiaanssen, University of Leeds – Job accessibility and employment outcomes: the role of public transport in the UK
Transport systems have an essential role in providing people with access to employment opportunities. This study examines the impact of public transport job accessibility on individual employment outcomes in the UK context. We developed a UK-wide public transport job accessibility measure, using novel public transport timetable data and detailed employment micro datasets following the widely used gravity model.
Jack Lewis, University of Leeds – Procedural environment generation for agent-based models of crime
Urban morphologies, such as patterns of land usage and street network configurations, fundamentally shape people’s day-to-day activities in complex and non-linear ways, dictating where, when, and in what contexts individuals interact with one another. As a result of difficulties associated with the empirical investigation of these interactions, scholars in recent years have applied agent-based models to study the relationship between urban morphology and space time behaviour.
This project aims to explore techniques for procedural environment generation for agent based models. Developing algorithms capable of generating realistic synthetic but experimentally manipulable simulation environments, we seek to explore how these techniques might increase our understanding of how street networks influence patterns of crime.
Stelios Theophanous, University of Leeds – Synergy PRIME – Multi level Modelling, Simulation and Visualization
Travel forecasting is at the core of urban transportation planning and plays a key role in determining the need for new road capacity, transit service changes and changes in land use policies. The Synergy PRIME project aims to integrate population growth projections, agent based modelling and interactive traffic simulation, while using immersive technologies and visualisation to test the potential future transport design scenarios. This novel approach to transport modelling could ultimately support the development of future intelligent transport systems.
Victoria Jenneson, University of Leeds – Supermarket purchase data in population dietary research
Population dietary surveillance is important for the epidemiological understanding of diet-related diseases and for targeting and evaluating public health policies. National dietary surveys rely upon self-reported intakes, which introduce recall and reporting biases and tend to underestimate intake.
This talk presents findings from a systematic review of electronic supermarket sales data in dietary research.
Francesca Pontin, University of Leeds – Physical inactivity; determining the demographic and geographic factors using smartphone data
Physical inactivity is a leading cause of non-communicable disease globally. Increasingly, individuals are using smartphone apps and wearable activity trackers which record physical activity behaviour, providing a new and novel source of data which are currently underutilised in research.
This study tracks the habitual activity patterns of over half a million people in the UK over a one year period, utilising physical activity data recorded by a smartphone app, linked to a range of wearable activity tracking technology. The temporal span of the dataset has allowed us to identify distinct seasonal trends in physical activity, for example, peaks and troughs in activity which coincide with UK daylight saving time changes.
Dr Michelle Morris, University of Leeds – What do you think about researchers using your lifestyle information: the LifeInfo Survey