DigitalExposome: A dataset for wellbeing classification using environmental air quality and human physiological data

Abstract

Urban environments play a critical role in shaping mental wellbeing, yet their impact remains understudied, particularly in relation to environmental air quality and human physiology. Despite this growing awareness of the importance of mental health in urban planning, challenges in integrating diverse datasets, spanning environmental, physiological, and self-reported mental wellbeing data limit the scope of research in this area. The DigitalExposome dataset addresses this gap by providing a comprehensive resource for understanding the relationship between these factors. The resulting data was collected from October 2021 to September 2022 in Nottingham, UK with the dataset including over 42, 437 samples from 40 participants aged between 18-50. Participants conducted a walk through diverse urban environments including polluted and green spaces, while carrying a custom-built environmental monitoring system (Enviro-IoT), wearing an Empatica E4 wearable, and using a smartphone mobile application to self-label mental wellbeing via emojis. Environmental variables (e.g., a range of particulates and gases including particulate matter and nitrogen dioxide), physiological metrics (e.g., HR, HRV, EDA, BVP), and mental wellbeing labels were recorded. Data was processed following collection through resampling and interpolation, and normalization for analysis. This novel dataset lays the groundwork for exploring the relationships between air quality, physiological changes, and mental wellbeing, offering valuable insights for urban planning and public health.