Melissa Clover: Communicating food sustainability: highlighting challenges and opportunities for eco-labels to enable informed consumer choices
Food sustainability has become a topic of increasing importance in recent decades, causing an increase in demand for communication of information between consumers and producers. This is commonly achieved via the medium of packaging in the form of eco-labels, which have become a popular tool for communicating sustainability information across a wide range of sectors in the last 30 years. Whilst the benefits of providing consumers with information are well-documented, the role of communicating sustainability within food systems requires further consideration. Additionally, with over 100 different types of registered food eco-labels worldwide, coupled with considerable evidence to suggest that existing modes of sustainability communication are leaving consumers confused, there is a need to examine existing types of eco-label. A review was therefore conducted to first understand how food sustainability information is communicated to consumers, before providing an overview of different types of eco-labels, categorised by a) international standardization offered by the UNEP, and b) different formats offered on packaging. This highlighted that communicating food sustainability is beneficial for both producers, who have the opportunity to improve consumer trust and monitor their environmental credentials in accordance with regulation, and consumers, who are empowered to make informed purchasing choices and have the potential to engage in more sustainable consumption. Additionally, the review highlighted the challenge of adaptability in eco-labelling, which is of increasing importance with instabilities in the global supply chain. The need for further research into the role of technology in addressing the need for more adaptable eco-labels is therefore highlighted.
Gregor Milligan: The Impact of Deprivation on Loneliness and Social Isolation
This paper explores the prevalence of loneliness and social isolation across London, UK and examines the co-occurrence of these phenomena in the context of economic deprivation indicators such as food and fuel insecurity. Through a comprehensive survey of loneliness, social isolation, and food and fuel security measures (n=2886), this work uncovers the relationship between characteristics such as deprivation, age, household composition and well-being on an individual’s propensity towards loneliness and social isolation.
Matthew Levesley: Critical Discourse Analysis of a website providing information about DIY HRT
Some transgender and nonbinary (trans) people will purchase hormone therapy without a prescription, in a process known as DIY HRT. This paper looks at an instructional website that describes how to purchase and take DIY HRT safely, in order to analyse the views on trans people, the medical profession, and HRT as a whole presented by the text.
Ruairi Blake: Personal and Situational Factors in Ethical Consumerism
12 UK Participants between the ages of 18 and 65 were interviewed on the topic of ethical consumerism, specifically their experiences and processes with the adoption of their lifestyle. The purpose of the study was twofold: Firstly, participants were asked to define ethical consumerism in order to create a more grounded definition of the term rather than using a top-down definition. Secondly, the responses to interview questions were analysed for personality and situational factors that might lead someone to adopt ethical consumerism as a lifestyle. In regard to a definition, the participants’ definitions mostly adhere to more general definitions used in academia, with the exception that focus is placed on intent rather than actual action. Thematic analysis showed that, in terms of personal traits, the most important were Knowledge-Seeking and Trying for an Ideal, though others were found. Situationally, participants reported either knowing someone else who is an ethical consumer or formal education as key factors in their decisions, among others. Other factors, such as choosing to try to ‘convert’ others were also found and discussed. These results fit within the general literature of motivations for ethical consumerism, giving further motivation to use these results as the basis of future studies.
Torran Semple: A Critique of the Low Income Low Energy Efficiency Approach to Measuring Fuel Poverty
This paper presents a critique of the Low Income Low Energy Efficiency (LILEE) fuel poverty metric, which is currently used in England. In part 1, a spatial analysis of Nottingham, UK, unveils potential discrepancies between deprivation and expected fuel poverty incidence. The number of financially vulnerable Energy Performance Certificate (EPC) A–C rated households in Nottingham is estimated to show that there is a considerable proportion of households currently classed as “not fuel poor” (according to the LILEE metric) that are still energy insecure (unable to afford sufficient energy). In part 2, survey data (n=2886) collected in London, UK, is analysed via the estimation of a random parameters ordered probit model. The model estimation results unveil the sociodemographic, behavioural and perceptual characteristics of respondents that significantly affected energy security (a condition closely related to fuel poverty) during the winter 2022 period. The sociodemographic characteristics that significantly increased the likelihood of energy insecurity were as follows: being aged 18–34, having a physical or mental health condition, household income less than £14,900 per year, and the presence of a prepayment meter in the home. The findings provide a conceptual framework for an amended or new fuel poverty metric in England, with a particular emphasis on the protection of vulnerable households who are more prone to energy insecurity.
Sam Smith: Automatic Lifestate Identification and Clustering
We present a novel approach to high-dimensional time series segmentation. The dataset contains 131 weeks of rail customer satisfaction survey data. We assume that each time-slice is drawn from a set of latent probability distributions (“lifestates”), then cluster all time-slices across the time series and calculate the latent probability distributions. This approach gives an insight into temporal changes and patterns in customer satisfaction data from the rail industry.
Carina Zhao: Using Intervals to Enhance Information Capture from Rail Passengers: An Exploratory Study
Railway transportation is crucial for connectivity across the UK, making it essential to collect insights from rail customers to promote industry development and growth.
Traditionally, the majority of data collection in this research context employs the use of discrete scales, such as Likert scales. This method requires participants to select a single value on a scale, potentially forcing them to make a choice even if they are uncertain, resulting in incomplete and biased feedback. This is particularly problematic given that the judgments and feedback provided by rail passengers are inherently subjective, and therefore contain uncertainty that may be of importance.
This paper investigates the potential of intervals in capturing uncertain information from rail passengers. Specifically, we adapt the Rail User Survey conducted by Transport Focus to include two different survey modes: one with discrete response options and another with Interval-Valued (IV) response options. By conducting the survey and collecting feedback from participants in a university environment, we demonstrate the potential and utility of the interval approach in obtaining richer information, especially uncertain information from rail passengers.
Our findings suggest that participants have varying preferences and attitudes for using interval or discrete response modes in different contexts. We argue that the existence of such preferences and attitudes justify the added value of intervals in this research. Therefore, with the data analysis of the current study still in progress, it is our contention that efforts should be made to further explore the reasons behind these preferences and attitudes.
Callum Berger: Towards a Novel Fear Model for Use Within Immersive Video Games
Through the synthesis of existing fear models in conjunction with the HCI trajectories model, this paper presents a new HCI model of fear and presence. This allows for an understanding of fear through a trajectory that can be applied within immersive video games.
Yang Bong: Predicting sharing platform growth: Using big data, network analysis and machine learning to forecast user acquisition
Hyperlocal sharing platforms have become increasingly popular as a way for people to access resources in a more environmentally-friendly and cost-effective manner. The success or failure of these networks depend entirely on local communities and the way they use the platform, but there is surprisingly limited empirical evidence on how these networks evolve over time and how the initial local starting conditions affect successful platform growth. In partnership with the world’s largest food sharing network, Olio, we examine 5,030,671 instances of food and resource sharing over 51 months in 312 districts across England. For each district we describe how local sharing networks develop over time and examine the attributes related to increased usage of the application. We subsequently develop a series of machine learning models to forecast network growth based on a range of engineered data sources, including: network analysis of user interactions, observed behavioural variation within the networks, and demographic and environmental characteristics of the local geographies in which each network emerged. A random forest classifier yielded best performance for prediction, which is subsequently analysed through SHapley Additive exPlanations variable importance measures to help explain what factors contribute to platform success and network growth. Finally, we identify four key managerial insights to inform strategies that can support network growth and foster the development of new sharing networks.
Gabrielle Hornshaw: Investigating the Challenges of Uncontrolled Hand Recognition
This paper explores the use of image recognition techniques on the 11k Hands dataset with the aim to identify and discuss some of the challenges of hand recognition. We train a series of models on different subsets of the data (right-dorsum, left-dorsum, right-palm, left-palm) and compare their performance on subset-wise and subject-wise ranked retrieval. We then train a series of models using Bayesian triplet loss to estimate a probability distribution for each embedding, and compare this to the traditional models in terms of ranked accuracy and calibration (association of output confidence with correctness). We find that treating each subset of the data as its own class when training outperforms using subject only. We also find some preliminary results that suggest variance in a probabilistic model can be useful when comparing images of different subsets, but these results are not conclusive. We finalise by suggesting ways to continue the research and combat the lack of data by utilising more hand-feature specific biometric datasets.
Nasser Alkhulaifi: Predicting Energy Consumption in Food Cold Storage Environments Using Multi-task Learning
The food and drinks manufacturing industry is a significant contributor to energy consumption and carbon dioxide emissions. Reducing energy usage and emissions can be achieved through better monitoring, aided by accurate prediction. Accurate predictions of energy consumption can lead to considerable cost savings for businesses and improved operational efficiency, as well as contribute to a more sustainable food system from an environmental perspective by minimising resource waste and carbon footprint. Nevertheless, achieving accurate predictions in food cold storage is challenging due to the complex interplay of numerous factors, including food type, quantity, size, packaging, ambient conditions, cooling system efficiency, operator activity, and door opening frequency, which together contribute to the intricate dynamics of energy consumption. In response to this challenge, this paper proposes a deep learning model based on a multi-task learning mechanism, which is particularly suited for capturing the intricate relationships among these interrelated factors. The effectiveness of the proposed approach is evaluated using real-world data collected from a food business in Nottingham, United Kingdom. The findings and comparative analysis demonstrate that the proposed model outperforms exist- ing methods and highlights the importance of considering human factors in energy consumption monitoring in a food cold storage environment.