Post #569: Virginia daily case counts to 3/24/2020, REVISED

Posted on March 24, 2020

Source:  Johns Hopkins Coronavirus Map, update through 9 PM EDT, 3/24/2020.  Historical data file for all days prior to that.

I have now figured  out, I think, what Hopkins changed about the data files, and they have released new ones.  At present, there is only an “official” time-series data set for the US as a whole, plus one-day cross-sectional counts for US states and counties.  This means that the final data on any of these graphs may be subject to some (hopefully small) revision when the next set of numbers comes out tomorrow.  This change in the data resulted in modest change in the historical day-by-day counts for the US, but nothing that mattered materially.

Hopkins has stopped tracking the number of people who have recovered, because they do not have (and never really did have) any hard data on that.  In addition, they plan to start listing the number of people tested, which will be a big help in analyzing differences in apparent mortality rates across countries.  (All other things equal, countries that test more and identify a larger number of mild cases should show lower overall mortality rate as a fraction of all identified cases.)

From the currently-released files, we can tell that Fairfax County has the most coronavirus cases in Virginia (46).  Here’s a text list of the current (9 PM 3/24/2020) counts for DC area.

Area Cases Deaths
Fairfax, Virginia, US 46 1
Arlington, Virginia, US 36 0
Prince William, Virginia, US 23 0
Loudoun, Virginia, US 18 0
Alexandria, Virginia, US 8 0
Stafford, Virginia, US 6 0
Fairfax City, Virginia, US 0 0
Falls Church, Virginia, US 0 0
District of Columbia,US 141 2
Montgomery, Maryland, US 107 1
Prince George’s, Maryland, US 63 2
Howard, Maryland, US 30 0
Anne Arundel, Maryland, US 24 0

Original writeup follows but is no longer relevant.

Caveat:  Hopkins changed the underlying data structure of its coronavirus database today.  It looks like they are transitioning the US data to a county-level basis, from states.  When something like that happens, you can’t quite be sure if today’s count is completely comparable to the historical data. 

Further, the Hopkins data files available to the public, as of today, no longer contain any US state data.  The accompanying discussion suggests that they plan to provide an additional file of detailed US data, but so far, that file is not available.

That said, up to this point, for Virginia, the 5 PM update of the map has almost always turned out to be, after the fact, the official count for the day.  (Presumably, any cases diagnosed after that point are counted in the next day’s total).

The upshot is that it appears that we had another day of (roughly) non-increasing additional case counts, per a hope expressed in Post #567.  But there’s some chance that’s an artifact of the change in the structure of the Hopkins database.

Upshot is, it remains too soon to tell whether Virginia has hit the inflection point.