Post #1112: COVID-19 trends to 4/18/2021

Posted on April 19, 2021

Is this the peak of the U.S. fourth wave?

Below is the U.S. and regional graph, in logs.  There really hasn’t been a trend to speak of for more than two weeks now.  But now, most of those regional graphs are starting to bend downward.

This past few days, all the places that were hotspots seemed to be getting a break.  The thing that started to look like a peak, for Michigan, a few days back, looks like even more of a peak now.  Minnesota seems to have gotten to a plateau.  All the other places where the U.K. variant is known to be highly prevalent are doing, well, pretty much nothing.  Like so:

Even Colorado, which was the outlier among Mountain states, and the Mountain state with the highest incidence of the U.K. variant, got little break:

The situation isn’t uniformly rosy.  Maine, which had largely been bypassed by COVID-19 in the earlier waves, now has a strong upward trend in new cases per day.  Both Oregon and Washington still have upward trends (but fairly low absolute rates of new cases per day).  Otherwise, you really have to look hard to find any dark clouds on the horizon.

Separately, vaccinations continue apace.  As I noted some time ago, these daily vaccination numbers have some issues.  In particular, they have a strong regular weekly cycle, so if you catch Sunday between your snapshots, it’ll look like the pace of vaccination has slowed.  (Which this next pair of snapshots does.)

I keep hearing news coverage about vaccination rates slowing down, and vaccine “hesitancy” and blah blah blah.  Sure as heck isn’t showing up in the national data.  I wonder of that’s a case of the news media cherry-picking a few places that make a good story.  But overlooking the actual U.S. trend.

Over the past three days, the elderly added another 0.33 percentage points per day, and the U.S. averaged about 3.7M vaccine shots per day.

Today’s data glitch:  I guess it had to happen sometime.  Month after month, states would find old cases and add then dump those into their case counts, generating a “speed bump” in their data.  On 4/17/2021, Missouri found over 7,000 old duplicate cases, and removed them from their case counts, per this documentation, creating, I guess, a “speed dip” in their data?  The problem is that  you can’t plot negative numbers on a log scale.  So I’m tossing those 7500 or so cases back into the Missouri data, and getting on with it.