The Association of Public Data Users held its annual conference in Arlington, Virginia, last week, featuring speakers from federal agencies—including the Census Bureau, the Bureau of Labor Statistics, and the USDA, state and municipal governments, academic research institutes, and more. Elizabeth Nash and I met with engaging data professionals who are creating novel solutions to persistent analysis problems.
Here are just a few of the themes that we heard discussed at the conference.
Insights from Administrative Data
Administrative data comes from records created for bureaucratic, rather than statistical, purposes. There’s an ongoing discussion of how to transform administrative data into datasets that can provide insights that survey data cannot. This has been very challenging.
When people establish a business, pay their taxes, or license a vehicle, the data that they enter on the necessary forms is stored in databases across the nation. These datasets can be assembled and standardized by government or private entities to create extremely large datasets that are then anonymized and aggregated. The LEHD employment data is an example of a dataset that the Census Bureau created using state unemployment insurance records.
The American Community Survey (ACS) is exploring possibility of supplementing survey data with tax assessment data to understand value of housing stock. Ashley Austin from the Census Bureau told us about a pilot project to use home sales records to supplement the home value question from the ACS. Because asking someone to assess how much they believe their home to be worth has produced notoriously inflated home value estimates, the ACS team is exploring ways to match administrative data with survey respondents’ households in order to achieve a more realistic understanding of housing value.
Patrice Norman from the Census Bureau told us about the Bureau’s strategy for consolidating the Survey of Business Owners, the Annual Survey of Entrepreneurs, and the Business Research and Development and Innovation Survey for Microbusinesses—three large costly efforts that have a lot of overlap—into a leaner Annual Business Survey, which they will supplement with administrative records.
The Census Bureau, 2020 and Beyond
The 2020 Census is right around the corner, and the big count was on everyone’s minds at the APDU conference. We heard from Enrique Lamas at the Census Bureau on their ongoing efforts to test the questionnaire and to hash out the logistics of conducting the nationwide count. On the outreach side, the Center for urban Research, CUNY Graduate Center, won a Data Viz Award for their “Census 2020 Hard to Count Online Map,” which shows census tracts with the lowest mail return rates from the 2010 Census.
And finally, we heard from Nancy Potok, Chief Statistician of the U.S., Erica Groshen, Chair of the Friends of the BLS, and John Thompson, Executive Director of the Council of Professional Associations on Federal Statistics (COPAFS) on the recent proposal to move the Bureau of Labor Statistics from the Department of Labor to the Commerce Department. Currently, the Commerce Department is home to two of the major national data agencies, the Census Bureau and the Bureau of Economic Analysis. According to the speakers, folding the Census Bureau in to the Commerce Department might make it easier for the statistical agencies to work together.
This was my first APDU conference, and it was one of those rare conferences where all the panels and speakers were engaging and dynamic. This post only hits a few of the highlights. If you’re a public data user, I’d recommend coming to the APDU conference next year.