Recreating geographical profiles from American FactFinder

On March 31, the Census Bureau announced that American FactFinder would be decommissioned, and this has created problems for those of us who used FactFinder to download data tables from the American Community Survey (ACS). Census has announced that the Geographic Comparison Tables (GCTs), which were formerly available on FactFinder, will not be available on the website. (The 2018 ACS one-year GCTs are available via FTP on; no ACS data before 2010 is available.)

So I’m wondering if I could use the SDA to create Table GCT0801 (Average Commuting Time to Work), using one-year ACS data from 2011, for the vintage-2003 metropolitan statistical areas. It looks like the IPUMS data have been harmonized so that data for the vintage-1993 MSAs can be retrieved - but it doesn’t look like I can create the same table for the MSAs with 2003 boundaries. If I try to create a county-level table for total minutes spent commuting to work, I get missing data for many counties, I guess because the county is not identifiable in the public-use data.

Sorry for the long question, but is it possible to use SDA to recreate Table GCT0801 for the one-year 2011 ACS data for MSAs with 2003 boundaries?

This sounds like a task that may be better suited for IPUMS NHGIS. In general, it will be difficult, and likely impossible, to entirely recalculate the information you are looking for using SDA with IPUMS USA data. This is due to the confidentiality restrictions placed on public use microdata that prevent relatively low sub-state level geographic identification.

IPUMS NHGIS uses the same source data as IPUMS USA (and American Fact Finder), but like American Fact Finder it provides tables with aggregated statistics. This allows for complete coverage of aggregated statistics at relatively low sub-state level geographical areas. Specifically, if you filter on “journey to work” as a topic and “2011” as a year you should find the information you are looking for.

This is terrific - thanks! I can’t quite reproduce GCT0801 (average time spent commuting to work) but I can get the number of workers in each category for total minutes spent commuting to work, so I can use those variables to estimate the average.