Thanks for the input. I have the shell and tables and even built my own hybrid table, so I can sort find, etc. I do this first. Then on a potential table to use, I download the column-metadata for it and choose all potential fields then pull the data from an API to see if there is data for that field (sometimes it's blank). I then have valid fields for that table (valid for the county level). I do this by table type (because the API format varies by type). I've spent 10-plus hours on subject tables so far this weekend. I then put all the valid subject table fields into one API call (max is 50 fields) so I can pull down all valid fields for the subject tables I want. B14005 does not have data at the county level. B15001 has valid data not included in the 25 and over (so it's a missing piece which is great). However, with thousands of tables, it's possible to build almost anything. Ed is a very basic and important measure there should be a field for it without searching for all the pieces to hopefully build the total measure. I think it should match enrollment that uses Age 3 or higher, and actually, there should be a top level with no age limit at all for both enrollment and attainment with breakout by age, race etc etc.
So.... I think Census data is very important and has a direct impact on many lives. Has some shortcomings which are worth fighting to get fixed. And in these cases, I think it's an easy fix as the raw data is already there. Trying to make the world a slightly better place, not just complaining. Offering no offense to anyone.