A Method for Designating Residential Socioeconomic Classes to Promote Community Security

Authors: Joseph Tuccillo*, Oak Ridge National Laboratory, Marie Urban, Oak Ridge National Laboratory
Topics: Population Geography, Urban and Regional Planning, Hazards and Vulnerability
Keywords: Census, Microdata, Housing, Housing Quality, Class, Socioeconomic Status, Demographics, Risk, Vulnerability
Session Type: Virtual Paper
Day: 4/10/2021
Start / End Time: 9:35 AM / 10:50 AM
Room: Virtual 41
Presentation File: No File Uploaded


A thorough inventory of housing stock is critical to promote human security in communities affected by emergencies or disasters. As a complement to physical characterizations of housing made via high-resolution imagery and remote sensing, it remains necessary to understand how residences differ compositionally based on high-resolution demographic and socially sensed data.

This research develops a semi-supervised classification approach to differentiate residential types by socioeconomic status, using census microdata from both domestic and international sources. We present a case study based on the American Housing Survey’s Metropolitan Public Use File (PUF) that separates single-family residences into “lower” and “middle/upper” classes for 10 metropolitan or “core-based” statistical areas (CBSAs) throughout the United States in 2019. Residential classes are distinguished based on the relationship between dwelling characteristics, including housing quality and material assets, and household income insecurity for the full range of housing types in the microdata. For each CBSA, this is achieved via a lower-dimensional embedding (redundancy analysis) of dwelling characteristics against occupant characteristics, weighted by household income insecurity.

To demonstrate how the residential class designations are applied in practice, we present examples of their use in informing daytime and nighttime dwelling occupancy estimates for Oak Ridge National Laboratory’s Population Density Tables (PDT) project. We conclude by exploring how such insights, in combination with methods for spatial refinement (i.e., population synthesis), can contribute to a more holistic understanding of community-level risk and vulnerability.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login