Journal of Data Science

High quality record linkages are critical for enriching survey data with alternative data sources. To enhance the American Community Survey (ACS) (US Census Bureau, 2025), we evaluated several options to match commercially available property data to housing unit records in the ACS and the Census Master Address File (MAF) (US Census Bureau, 2022) data, with the ultimate goal of supplementing data collected in the survey with the commercial property data. Techniques to match address data come in two flavors: spatial matching and address matching. Spatial matching is done by overlaying commercial boundary shape files on the lat-long coordinates on the MAF to associate Census housing unit records to commercial property parcel records. This method is useful because it does not require matching of text fields, but performs poorly when parcels include many housing units (e.g., large apartment buildings). Address matching, or entity resolution at the address level, links records across the two sources based on the content of various address fields, and offers the possibility of disambiguating multiple matches and matching objects that cannot be successfully assigned a unique match through spatial matching. Several approaches are available for address matching, including rule-based, deterministic, fuzzy, and probabilistic methods. In our article we summarize literature comparing various combinations of spatial and address matching techniques to illustrate the trade-off between linkage rates and linkage quality. We also consider hybrid solutions that leverage spatial matching as well as several types of address matching to maximize high quality linkages for our research and discuss future directions such as incorporating probabilistic matching as an added step.

The United States has a racial homeownership gap due to a legacy of historic inequality and discriminatory policies, but factors that contribute to the racial disparity in homeownership rates between White Americans and people of color have not been fully characterized. In order to alleviate this issue, policymakers need a better understanding of how risk factors affect the homeownership rates of racial and ethnic groups differently. In this study, data from several publicly available surveys, including the American Community Survey and United States Census, were leveraged in combination with statistical learning models to investigate potential factors related to homeownership rates across racial and ethnic categories, with a focus on how risk factors vary by race or ethnicity. Our models indicated that job availability for specific demographics, and specific regions of the United States were factors that affect homeownership rates in Black, Hispanic, and Asian populations in different ways. Based on the results of this study, it is recommended policymakers promote strategies to increase access to jobs for people of color (POC), such as vocational training and programs to reduce implicit bias in hiring practices. These interventions could ultimately increase homeownership rates for POC and be a step toward reducing the racial wealth gap.

Racial and ethnic representation in home ownership rates is an important public policy topic for addressing inequality within society. Although more than half of the households in the US are owned, rather than rented, the representation of home ownership is unequal among different racial and ethnic groups. Here we analyze the US Census Bureau’s American Community Survey data to conduct an exploratory and statistical analysis of home ownership in the US, and find sociodemographic factors that are associated with differences in home ownership rates. We use binomial and beta-binomial generalized linear models (GLMs) with 2020 county-level data to model the home ownership rate, and fit the beta-binomial models with Bayesian estimation. We determine that race/ethnic group, geographic region, and income all have significant associations with the home ownership rate. To make the data and results accessible to the public, we develop an Shiny web application in R with exploratory plots and model predictions.

The 2020 Census County Assessment Tool was developed to assist decennial census data users in identifying deviations between expected census counts and the released counts across population and housing indicators. The tool also offers contextual data for each county on factors which could have contributed to census collection issues, such as self-response rates and COVID-19 infection rates. The tool compiles this information into a downloadable report and includes additional local data sources relevant to the data collection process and experts to seek more assistance.