Many Americans live in two distinct community forms: (1) informal subdivisions (ISs), where residents use incremental self-building for housing development that does not adhere to formal land planning and housing construction practices and (2) manufactured home communities (MHCs), where the dominant housing model is manufactured, low-cost factory-built housing that provides the single largest source of unsubsidized affordable housing in the United States. Although these communities provide a major source of affordable housing and low-income home ownership, case study research suggests that they are spatially marginalized and exposed to concentrated forms of economic, social, and environmental vulnerability. Due to the difficulty of identifying their location across a broader geography, there is currently no systematic data on their total number or location, nor are there national-level analysis of the spatial inequalities they face. This project uses big data and machine learning to produce more robust and refined measurements of the characteristics of all U.S. neighborhoods (formally planned suburbs, ISs, and MHCs). This allows documentation of the location of ISs and MHCs nationwide and modeling of policy and market factors that explain patterns of uneven development, segregation, and environmental inequalities across neighborhood types. The databases and publications generated by this project have the potential to generate knowledge needed to develop more equitable housing policies as well as to support further research. The dissemination plan allows knowledge sharing with the public, local planners, and other stakeholders and policymakers through local community engagement workshops, a series of regional webinars, and an easy-to-use publicly available data mapping and visualization dashboard.
The study builds on geographic theories of socio-spatial peripheralization and uneven development by examining the nature, causes, and consequences of the proliferation of ISs and MHCs and their relationship with the uneven spatial distribution of poverty and vulnerability in the United States. The project uses Python programming language, a national dataset of building footprints, and supervised and unsupervised machine learning methods to identify the distinct dimensions of neighborhood morphology (the size, shape, orientation, and other arrangements of buildings) in ISs, MHCs, and formally planned suburbs across the country. In doing so, it produces more robust and refined measurements of the characteristics of all U.S. neighborhoods, as well as a first-time national level database of ISs and MHCs. This dataset enables the examination of the relationship between segregation by neighborhood types and spatial inequalities, including residential segregation by race, income, and tenure as well as exposure to various types of environmental risk. Project findings contributes to: (1) methodological advancements in the spatial study of neighborhood morphologies, (2) theoretical advancements in scholarship on peripheralization, uneven development, and suburbanization of poverty and (3) empirical advancements in the documentation and analysis of informal housing relative to social vulnerabilities and environmental hazards. The study allows users of the research products to analyze neighborhood morphologies; examine social, economic, and environmental impacts of uneven community development; and identify policies that can ameliorate these impacts.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
