Hello, I have a Python process that gets values for all tracts in the US for a specified table (it drops water tracts, but other than that keeps everything). I ran it on the 2010-2014 data last fall for table B15002 and got 73,638 records. I ran the same, unchanged, process again yesterday and today, 2010-2014 data for B15002, and this time got 73,570 records. What's even stranger is it's not just dropping records, there are some new GEOIDs that weren't there last time (counties 02275 and 46099), and there are also some missing GEOIDs compared to last time (counties 23005 and 27105). Has there been a change to the ACS API for the 2010-2014 estimates since last year, esp. in terms of what counties come up?
Thanks,
Diana