Below are links to data files I produced that show the number of new COVID-19 cases *per day* for each county in the USA. These are .csv files and also include coordinates for each county centroid. This means the files are easy to import into spreadsheet or GIS software. The only limitation is that your use must be non-commercial. This data will be updated each night as the pandemic continues.
The source of the data is a Johns Hopkins University GitHub page at:
https://github.com/
A team at Hopkins collects information from a variety of sources and publishes a daily *cumulative* count of COVID cases for each county. This data is widely regarded as authoritative. If you are curious, here are some links where you can learn more about this work.
A few weeks ago I decided it would be useful if the Hopkins cumulative count data was converted to show the number of new COVID19 cases per day for each county. After searching online and failing to find the Hopkins data converted to daily counts, I decided to take on the task of writing code to do that conversion.
That code now runs automatically each night (12:10a.m. pacific time), reads the prior two days worth of data from the Hopkins GitHib site, computes the new cases for yesterday and adds a new column to the csv file for the current month. Each month a new csv file will start. The daily count data starts on March 24. (Prior to March 23 there was a period of time when Hopkins was only reporting state totals.)
Below is a link to the March csv file. The same naming convention will be followed as new months and years come along. In addition to this March file, you can also presently download a csv file for April, May and June by changing the month number in the following link.
March csv file:
https://mappingsupport.com/p2/
Yes, the number of new COVID19 cases in these csv files might be different than the case numbers reported by county health departments. There are various reasons for any differences. For example, Hopkins counts both confirmed and probable cases. Also Hopkins ‘scrapes’ data from state health department websites and that data might lag a day or two behind data on county websites.
You also will sometimes see negative numbers in these csv files. That might indicate a person who was counted as probable but then removed from the count when a negative test result was returned in a few days. Or maybe a state or county changed its definition of what constitutes a case.
Note that the Hopkins GitHub site actually has 2 sets of data. Each “daily” file holds the cumulative count of cases for all counties for a single day. These are the files my code presently uses to compute the number of new cases per day. The other set of data is a “timeseries” file of new cases. Each day that file gets bigger as the cumulative counts of cases for all counties for the prior day is added to that file. Turns out that Hopkins sometimes updates the data in the timeseries file but does not update the daily files. Alas, this important tidbit is not mentioned in the README for the daily files. Hopefully by the end of June I will have converted my code to use the Hopkins timeseries file.
If you use these csv files please credit both the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University and Joseph Elfelt, MappingSupport.com (linked to https://mappingsupport.com)
Finally, I am hoping that this information finds its way to teachers at the middle school and high school level. Students can use this data to make their own charts and graphs showing what the pandemic is doing per day in their county and state. And just maybe students can help ‘move the needle’ by convincing their parents to wear a face covering and practice social distancing.
Joseph Elfelt
Redmond, WA
Twitter: @mappingsupport
https://mappingsupport.com
6/18/2020