Image credit: Unsplash

PAA 2020: The US Census, Differential Privacy, & Geographic Analysis


The US Census Bureau has growing concern that with the rise of publicly available data on individuals, that information from the census may be at risk for re-identification. As a preemptive measure to mitigate the risk of re-identification the bureau is moving forward with implementing a version of epsilon differential privacy for the 2020 Census. How this will alter geographic analysis has yet to be discussed and in this talk we open the discussion of geographic analysis and differential privacy with a discussion on how geographic statistics of concern for social scientists are altered in the 2010 Census end to end test.

Population Association of America Conference
Seattle, Washington
Neal Marquez
Forecast Program Manager

My research interests include demographics of Hispanics in the United States, contextual influences on population processes, validation of digital trace data, and spatio-temporal bayesian modeling matter.