When the New York Times made its compilation of covid cases available, I decided to take a look. Although it seems like EVERYONE is doing covid analysis, there were a few things that I wasn’t seeing. For instance, the US is a huge country and I imagine that the sense of threat from covid is going to be related to how close the virus is geographically to a given person.
At the state level, at the beginning of March, only about 10 states were showing cases, but within a week or so, almost all states registered cases. By the beginning of April, just about all states registered a covid death:
Because the NYT data are reported down to the county level, we can get better geographic resolution. At this level, we see there are still quite a few counties that haven’t registered a case and less than half have seen a covid death (As of May 6, 2020).
How about looking at the number of people in those counties? For that, I pulled in the 2010 Census numbers for a quick comparison (I removed territories for this analysis). The total population for 2010 is 309,585,169. Here we see that just about all people are in counties with cases and about 85% of people are in counties that have registered deaths – many of those ‘spared’ counties have very few people.
And here are the days of first cases and deaths:
(Here is a version of the Jupyter notebook I used for these analyses.)