Title:
Assessing mental wellbeing in urban areas using social media data: understanding when and where urbanites stress and de-stress

Thumbnail Image
Author(s)
Dutt, Florina
Authors
Advisor(s)
Guhathakurta, Subhrajit
Advisor(s)
Editor(s)
Associated Organization(s)
Organizational Unit
Supplementary to
Abstract
Are Americans more stressed out by living in dense, urbanized areas or less dense, car-oriented areas? To answer this question, can we use people's expressions of stress in different environments to understand what kinds of spaces help them de-stress? This study uses stress levels of geolocated tweets to help us answer such inquiries and resolve the longstanding disparities between the field of psychology and urban planning about mental health impacts of cities. This is important because more than 75 percent of Americans are moderately stressed. Long-term stress is associated with mental health disorders, including sleeplessness, anxiety, and depression. Additionally, chronic stress is linked to physical ailments, including high blood pressure, cardiovascular diseases, and diabetes. The psychology literature claims that urban areas witness elevated levels of mental health problems, manifested as stress, mood disorders, and anxiety issues. Density, crowding, traffic, crime, and pollution are identified as stressors associated with urban living conditions. Contending this claim, the urban planning literature positions stress in the context of longer commutes, lack of accessibility, and social isolation that comes with suburban living conditions. Urban Planners and urban designers have advocated for density. With rapid urbanization, 60 percent of the world population will live in urban areas by 2030, making it crucial for urban planners to address these disparities to support the mental wellbeing of the urbanites. This research uses multi-headed attention transformer model to classify tweets (token sequences), and assesses the stress levels of custom-defined assessment grids of ten acres within the city area of Atlanta and Boston. The assessed stress level of these assessment grids is called the mental wellbeing score (MWS). Mental wellbeing score is defined in this research as a measure of `mental wellbeing' of any given grid (higher score is better). Using this measure, the research investigates the relationship between mental wellbeing and built environment characteristics in urban areas to uncover the impact of long-term stress triggered by the conditions of the built environment in urban settings. In summary, the results of the exploration shed light on three critical aspects: 1. Mental wellbeing score increases with increasing urbanness. 2. The mental wellbeing score increases with the increase in the diversity of escape facilities, including green parks, open spaces, and other points of interest. 3. The mental wellbeing score is positively impacted by accessible high-density spaces with high symbolic value. The research also investigates the impact of safety perception and socio-economic status on mental wellbeing scores. The results show that addressing socio-economic disparity, crime, and investment in green infrastructure can improve mental wellbeing of urbanites. The methods and findings of the research show that 'urban areas' can positively impact mental health if designed appropriately. Furthermore, this study can empower urban planners and policymakers to develop tools to assess the mental wellbeing of urbanites, adjust infrastructure needs, and improve the urban amenities that support mental wellbeing.
Sponsor
Date Issued
2021-12-10
Extent
Resource Type
Text
Resource Subtype
Dissertation
Rights Statement
Rights URI