While detailed geographic information is an important component of data utility, sample size and privacy concerns can make release of small area statistics (e.g., census tract statistics) challenging. To address these issues, NORC is working to develop estimation methods that generate modeled estimates at lower levels of geography than are available in public use microdata sets, allowing for a much more detailed picture of person and housing characteristics than would have been otherwise possible.
As one example of the use of these methods, NORC created a visualization tool that shows modeled (census) tract-level measures of housing characteristics from the New York City Housing Vacancy Survey (NYCHVS). We also plotted how these relate to a measure of gentrification taken from tract-level American Community Survey data so that users can examine how housing characteristics relate to gentrification. This tool won the Joint Statistical Meeting Data Expo 2019 prize in the professional category.
This interactive tool won the 2019 American Statistical Association Joint Statistical Meetings Data Challenge. It was designed and created by Quentin Brummet, Peter Herman, Evan Nathan-Herring, Edward Mulrow, and Nola du Toit.