Data
Every series we publish, as a CSV. The URLs are stable and are meant to be cited;
they serve Access-Control-Allow-Origin: *, so you can
fetch them straight from a browser or a notebook.
Each file is long format, one row per observation:
fips,county,period_start,value,flag,vintage
A dataset split into sub-series — stations, industry sectors — carries a
series_key column after county.
period_start is the first day of the period, whatever its
length. value is empty where the source published
nothing or withheld it, and flag says which;
an empty value is never a zero. vintage is the date
that figure was published, so a number you cite today stays identifiable after
the source revises it.
Nevada
Vintages differ by dataset because each source publishes on its own schedule; each dataset page states the one it is showing.