Related Experiment Video
Updated: Aug 5, 2025
08:25
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
Published on: May 7, 2019
9.0K
DigitalExposome: quantifying impact of urban environment on wellbeing using sensor fusion and deep learning
Thomas Johnson1, Eiman Kanjo1, Kieran Woodward1
1Department of Computer Science, Nottingham Trent University, Nottingham, UK.
Computational Urban Science
|March 27, 2023
Summary
Environmental air pollutants impact mental wellbeing. This study introduces the
Area of Science:
- Environmental Health Science
- Digital Health
- Human-Computer Interaction
Background:
- Rising air pollution levels (particulates, noise, gases) negatively affect mental wellbeing.
- Existing research lacks a comprehensive framework to link environmental exposures with personal characteristics, behavior, and wellbeing.
- Mobile sensing technology offers potential for real-world data collection on environmental exposures and physiological responses.
Purpose of the Study:
- To introduce and define the 'DigitalExposome' conceptual framework.
- To investigate the relationship between urban environmental factors, physiological reactions, and perceived wellbeing.
- To utilize multimodal mobile sensing for simultaneous data collection in urban settings.
Main Methods:
- Simultaneous collection of multi-sensor data: urban environmental factors (air pollution, noise, people count), physiological reactions (Electrodermal Activity (EDA), Heart Rate (HR), Heart Rate Variability (HRV), Body Temperature, Blood Volume Pulse (BVP), movement), and self-reported wellbeing (valence).
- Data collection via a comprehensive sensing edge device during a pre-specified urban path.
- Multivariate statistical analyses (Principle Component Analysis, Regression, Spatial Visualizations) and Convolutional Neural Network (CNN) for data analysis and wellbeing classification.
Main Results:
- Electrodermal Activity (EDA) and Heart Rate Variability (HRV) were significantly impacted by ambient Particulate Matter levels.
- A Convolutional Neural Network (CNN) model achieved an f1-score of 0.76 in classifying self-reported wellbeing from multimodal sensor data.
- The study demonstrated the feasibility of collecting and fusing multimodal sensor data in real-world urban environments.
Conclusions:
- The 'DigitalExposome' framework provides a novel approach to understanding the complex interplay between environment and wellbeing.
- Objective physiological measures (EDA, HRV) are sensitive indicators of exposure to air pollution.
- Multimodal sensing and machine learning offer promising avenues for monitoring and potentially improving mental wellbeing in urban populations.
