Data appears heterogenous; some is not up to date

[ballcoach12]

It is unclear from reviewing the Confirmed Cases data set whether the United States data is being tracked only by state or my state/county. There are entries for both, and it isn't clear which one should be used for modeling. Also, data for Hamilton County, TN is inaccurate. Three cases have been confirmed and are not in the file according to county/state. I assume that these cases are in the state-only tally.

[yetzt]

See #1054

[apsteinmetz]

Agreed. Is the local data disaggregated state data so using local and state level is double counting?

[Sleekery]

After 3/10/20, they started keeping track in the US by state only, no county/city level. You'll have to aggregate pre-3/10/20 data into states first. Hint: they're the only values in Province/State with commas. Filter by that, split by commas, and then map US postal codes to states.