Unveiling the Wall: Physical Manifestations of Socio-spatial Boundaries in the US
Open Access
- Author:
- Rahimi, Sohrab
- Graduate Program:
- Architecture
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- April 15, 2019
- Committee Members:
- Mallika Bose, Dissertation Advisor/Co-Advisor
Mallika Bose, Committee Chair/Co-Chair
Clio Maria Andris, Committee Member
Xiang Zhang, Committee Member
Jose M Pinto Duarte, Outside Member - Keywords:
- Urban Sociology
Urban Culture
GIS
Machine Learning
Social Media
Yelp
Urban Computation
Taste
Bourdieu
Social Boundaries
Segregation
American Cities - Abstract:
- This research aims to identify and characterize the socio-spatial boundaries that separate urban communities in American contemporary cities. The American contemporary city, also known as the postmodern city, is characterized by high social mobility and extreme social polarization. Studying these cities with highly dynamic socio-spatial boundaries is challenging and conventional data sources such as US Census fail to provide a full picture of these boundaries. As an alternative method, this study uses a crowd-sourced social media data (Yelp) to identify these socio-spatial boundaries in the context of postmodern cities. By drawing on Bourdieu’s theory which asserts that taste is an indicator of identifying social groups, I identify urban communities in different cities through analysis of Yelp reviews. I also investigate the physical characteristics of the sociospatial boundaries between these communities. The first part of this study validates Bourdieu’s theory of distinction in the context of American cities. I use Natural Language Processing (NLP) techniques to see if the frequency of certain food and drink types as mentioned by Yelp reviews in different regions can be a good proxy of social class, as defined by Bourdieu. Through statistical analyses, I show that taste for food and drink is an acceptable proxy for social class which is conventionally characterized by education, income and race. While I investigate the association between food and drink types discussed in the Yelp reviews (by restaurant patrons) and demographics of neighborhoods in which the restaurants are located I recognize that there are no means to ascertain if the patrons of the restaurants are residents of the neighborhood in which the restaurants are located. Furthermore, I use taste as a way to distinguish between different groups and not as a means to categorize groups in a hierarchical manner. In the second part of the study, I investigate the extent to which observable spatial patterns can be found by using taste as an indicator. The first part of the study (discussed in the previous paragraph) validates Bourdieu’s theory however, it does not show whether or not taste is capable of showing geographic/spatial patterns and can be used as an indicator for identifying spatially located urban groups. In this part, I first identify food and drink types that distinguish Yelp users in their ratings or choice of restaurant. I then use these selected set of foods and drinks to find the social groups by using unsupervised techniques. Our clustering results showed clear spatial boundaries between groups of people with different taste in food and drink. In the third part of this study I investigate the socio-spatial boundaries which I found in the second part of the study. Based on Lefebvre’s theory, I argue that there are two possibilities at the border lines: boundaries or junction points. Boundaries separate one neighborhood from another, while junction points connect the communities. Accordingly, I train multiple classifiers to see which urban and physical factors, as extracted through an extensive literature review, are associated with formation of these junction points and boundaries. Our findings are consistent with previous iv studies. I found that junction points have higher density of businesses, are more walkable, and are less likely to have high percentage of service industries such as warehouses and transportation related. In this tripartite study, I showed that Bourdieu’s theory of distinction is still valid in the context of the US. I also showed that socio-spatial boundaries can be identified through analyzing social media datasets that are more current, have higher spatial resolution, and lower costs compared to the Census data. Lastly, I showed that the socio-spatial boundaries identified through this process, have certain physical characteristics which can help urban planners and policy makers to make more informed decisions. Our methods, therefore, puts forth a new method to understand the increasingly complex social dynamics of American cities and gives new directions to future research to further investigate the urban boundaries using different other datasets and statistical methods. The results of this study can help planners, geographers and social scientists gain a better understanding of the socio-spatial structure of contemporary cities and more importantly, design friendly cities that are inclusive to all social groups.