Some counties have decreasing cases over time

[charlie86]

Hi, thank you so much for releasing this dataset. I noticed a few cases where a county has fewer cases than on previous days: here is an example with Kittitas, WA. Is this an issue from the county reporting or an intentional correction?

source: https://rcharlie.shinyapps.io/covid-19-data
code: https://github.com/charlie86/covid-dashboard

[albertsun]

@charlie86 cases where the number of cases drop from one day to the next are when the county or state has reassigned cases from one county to another. Or when they've discovered the cases were previously incorrectly reported.

If we aren't able to find more detailed information to correct the previous count, we'll simply report the counties totals dropping like that.

[charlie86]

@albertsun Got it, thank you!

[Adam-Hoelscher]

@albertsun Are you actively following up with officials in those places? Would your team find it valuable to have a list of county/date combinations where this occurs (for that reason or others)?

[albertsun]

@Adam-Hoelscher thanks for your kind offer. We're already reviewing the list of places it's happening following up with health officials and other data providers when warranted.

[ke4roh]

https://www.co.kittitas.wa.us/press/default.aspx?prID=11364

Kittitas County, WA - 03/25/2020 - Kittitas County has six cases of COVID-19 as of March 25th, 2020. The information on the Washington State Department of Health (DOH) website had an error listed in the case count, which has since been updated. Kittitas County does not have 18 confirmed cases of COVID-19 at this time. The Washington State Department of Health immediately corrected their error as soon as we notified them.

This was not an error with local data. The Kittitas County Public Health Department (KCPHD) is notified of every positive case in our county. We provide that information, immediately, to the Washington State Department of Health. “This was a unique and unfortunate error,” states Health Officer Dr. Mark Larson. “DOH was responsive and their site is still where we want people to go to for accurate information.”

[ke4roh]

Fixed.

[NickSto]

Not sure if you mean it's been fixed in this dataset, but just FYI this issue is still present in at least one other instance:

date	state	fips	cases	deaths
2020-04-06	District of Columbia	11	1097	24
2020-04-07	District of Columbia	11	1211	22

[ke4roh]

A complete list of these decreases is at https://github.com/ke4roh/covid-19-data/blob/cumulative_negatives/negativeCountyIncrements.csv