Exploratory Analysis of Spatial Hierarchical Clustering in Los Angeles County,California: Relationship of Opportunity-Based Accessibility,Reported Land Values,and Resident Characteristics

Abstract

The phenomenon of agglomeration has long been recognized as an important component in businesses and their location decisions. A similar phenomenon may also exist in resident location of the population in an urban area. This concept may also emerge from residential segregation and ethnic enclaves. Residential segregation may be influenced by location characteristics, social composition, and accessibility. This paper uses spatial clustering indicators that comprise resident population and its characteristics, infrastructure provision, activity opportunities by opportunity type, housing supply, and synoptic measures of travel. Time of day is introduced as a fundamental unit in classifying the spatial units. Classification of spatial units is accomplished first with a cluster technique that accounts for spatial dependence in opportunity-based accessibility indicators for different variables considered individually. Normalized scores for each variable and all spatial units are then used to derive a second set of clusters using multiple variables to build a group of clusters of spatial units. The characteristics of the residents in each spatial unit are examined for the Los Angeles County region of California with a sample of approximately 6,000 geolocated households with corresponding data on their residences along with detailed location characteristics. Residents of these clusters share similarity in their characteristics, and this analysis provides an example not only of a new way to use spatial clustering to understand location choices of households but also a way to alleviate the problem of dimensionality in choice set identification.