Life satisfaction significantly contributes to wellbeing and is linked to positive outcomes for individual people and society more broadly. However, previous research demonstrates that many factors contribute to the life satisfaction of an individual person, including: demography, socioeconomic status, health, deprivation, family life, friendships, social networks, living environment, and the broad range of behaviours enacted by the person, such as helping or volunteering. Consequently, it is challenging to disentangle the factors that contribute most significantly to life satisfaction, and thus more importantly, inform public policies designed to help foster positive wellbeing. We analyse primary survey data $(\mathrm{n}=2849)$ on self-reported life satisfaction in relation to a range of self-reported and observed variables associated with wellbeing. Specifically, we draw on a massive paired dataset related to use of a food sharing application in London, to augment the analysis using additional socioeconomic, environmental, and behavioural variables. Through a random forest machine learning approach and variable importance measures, we evaluate how a range of factors, that are often only evaluated individually, provide relative contributions towards life satisfaction. Result reveal that factors such as employment and social reliance contribute most significantly towards the experience of life satisfaction.

Big data from food retail stores is increasingly being used for population dietary surveillance, epidemiological studies of diet-related diseases, and evaluations of public health interventions. However, for retail data to be useful it is necessary to understand the spatio-temporal variation of when and where food is purchased and consumed. While some customers willingly share home location data with retailers as part of loyalty programs such data is typically too fine-grained/sensitive to be applied for research purposes. The aim of this study was to analyse differences between privacy-preserving models and actual retail catchments, and investigate if machine learning techniques could improve the accuracy of such catchment models. Based on a UK-wide sample of 4 million grocery store loyalty card holders, covering 485 million transactions over 29 months (2019-2021) and distributed across 33,000 neighbourhoods (Lower Super Output Areas, or LSOA), the study demonstrates how models trained on geolocated data perform at predicting, per store, catchment areas which contain 50, 80, and 95% of its customers’ primary location. Through comparative assessment of machine learning approaches, we find better performance from tree-based models (RF, XGB) with the best performance from an XGB model achieving an R2of 0.72 and MAE of 1.06. To conclude, we review variable importance measures using SHAP values and discuss the relative merits of including specific features when modeling catchment areas.

The redistribution of surplus food is a challenging problem, yet a crucial one to address given the urgent nature of climate change. However, designing computer-mediated food sharing systems is made even harder due to failed interactions between users and ensuing complaints, which can dissuade others from participating when shared within a public forum. To examine the phenomenon of complaints within such data, we analyze the public forum of a food sharing platform, OLIO. We characterize complaining behaviour and augment it through qualitative labeling and a machine learning approach to model complaints using affective indicators of dissatisfaction across a corpus of 3,195 forum posts. Results emphasize that linguistic features yield high prediction accuracies, with negative, nonconstructive sentiment being of greatest relevance. We discuss how machine learning can further enrich qualitative understandings and validation of complaints in the sharing economy.