Spatial–Temporal Neighborhood Patterns in Four Legacy Cities

Abstract

Legacy cities are characterized by long-term, declining trends in both population and economic characteristics, but how these events translate to the neighborhood scale is less well understood. This research investigates the evolution of neighborhood types in four legacy cities—Baltimore, Cleveland, Philadelphia, and St. Louis—from 1970 to 2010. Working from a multidimensional framework of variables across five census decades, hierarchical cluster analysis and discriminant analysis are used to develop a neighborhood typology, identify temporal trends, and describe the pathways of transition. Results reveal four crosscutting neighborhood types: Black Distressed; Lower Middle; Multifamily, Educated, Turnover; and Upper Middle, which persist across space and time. Most neighborhoods (61%) remained classified within the same type over 40 years, and transitions were concentrated in two decades (1970–1980 and 2000–2010). The results offer a new descriptive dimension for understanding the unevenness present across legacy cities and suggest that existing policy approaches continually reproduce the same neighborhood outcomes.

Changing neighborhoods, shifting connections: mapping relational geographies of gentrification using social media data

“…On the surface, the patterns of mobility and visits of Twitter users remain relatively stable over the five-plus years of the study period. This echoes findings elsewhere in the literature that neighborhoods mostly stay the same over time rather than experiencing drastic change (Solari 2012;Cortright and Mahmoudi 2014;Delmelle 2017;Malone and Redfearn 2018;Connor et al 2020;Kinahan, forthcoming). In this case, however, our analysis shows that this stability in neighborhood character not only applies to the internal characteristics of neighborhoods, but also how neighborhoods connect to each other relationally.…”

Neighbourhood change in Genesee and Kent Counties, Michigan, 1970–2019

“…They suggest the geodemographic approach that uses clustering algorithms to group census tracts into larger entities and allows tracts to form new partnerships with contiguous tracts as characteristics of residents change over time. However, I am interested in comparing neighbourhood change results with other studies on neighbourhood change in “Rustbelt” cities (e.g., Fee, 2017; Kinahan, 2021; Li & Xie, 2018; Mikelbank, 2011; Morenoff & Tienda, 1997) that have used census tracts as the unit of analysis and therefore do not adopt Rey et al (2011) methodology.…”

“…After reviewing several studies on neighbourhood change (e.g., Baum‐Snow & Hartley, 2020; Delmelle, 2017; Kinahan, 2021; Wei & Knox, 2014), 14 variables were selected to represent neighbourhood characteristics. These included demographic variables (percentage Black, percentage Hispanic, and percentage population over age 60); housing variables (median housing value, vacancy rate, percentage homeownership, percent of housing that is multiunit, percentage of households composed of families, percentage of population in neighbourhood less than 10 years) and socio‐economic variables (median household income, percentage in poverty, percentage of population over 25 with college degree, and percentage of labour force in manufacturing).…”

“…All data were processed in the statistical program SPSS. Each variable for both counties was transformed to z‐scores to allow comparison over time which is a process used by other neighbourhood change researchers (Delmelle, 2017; Kinahan, 2021; Wei & Knox, 2014).…”

“…Several scholars have opted for a third approach (the comparative approach) to neighbourhood change studies. These scholars examined between 2–4 cities and noted similarities and differences between the cities in reference to neighbourhood change over several decades (Delmelle, 2016; Fee, 2017; Foote & Walter, 2017; Kinahan, 2021; Wei & Knox, 2015). Regardless of the approach taken, each of the studies found that stability (defined as at least 50% of neighbourhoods maintaining their status over the study period) in neighbourhood type was the norm.…”