Thread by @dhmontgomery

#COVID19 data is going to be very messy the next week or two because of Thanksgiving. Not only did the holiday mess up when data is REPORTED, it also messed up how the data is collected. Ordinarily MN has tens of thousands of new COVID tests conducted on a Thursday….

Also @mnhealth isn’t reporting any vaccination data today. We’ll see how my scripts handle that…

Case in point of the weird data: today’s MN #COVID19 report has more than 104,000 total tests. That’s double what we reported last Monday, but of course the holiday means it’s not apples-to-apples.

That spike in reported tests means MN has a freakishly low daily positivity rate today — 4.3%, the lowest since August, which in turn drives our average positivity rate from 9.2% to 7.6%. That’s HUGE — and probably partially artificial. Maybe expect a bounceback in near future.

Newly reported cases are down about 4.4% week-over-week, which is PROBABLY a better gauge for where things are going based on pre-holiday trends. My 7-day average calculations are showing a much bigger ~18% dip because we’re missing a day, but I’d be cautious there.

Cases by SAMPLE date will be more robust this next week, but unfortunately they have a one-week lag. Last Monday’s tests had about 5,000 cases reported, a week-over-week drop, despite increased pre-holiday testing.

Lots of my graphs show misleadingly large drops today due to the holiday, so I’m sharing them sparingly until we get a better idea what’s going on.

Our pre-holiday trend was for cases trending flat or slightly down (with the possibility we had peaked a bit before Thanksgiving).

That flat-or-slightly-down trend is going to be my prior assumption for what’s going on in this week of messy data.

Regardless of what’s happening, one thing is clear: it’s been nearly 2 weeks since MN last had a significant upward spike in case counts. That’s *really* encouraging, after weeks when numbers soared upward every day.

In general, I am feeling cautiously optimistic.

We COULD be about to see a major drop in cases & other metrics in MN. It’s just too early to say for sure.

We could also see a long-term drop in cases, temporarily interrupted by a short post-holiday bump.

We could also see cases sputter on along a plateau for a while yet.

This could even be yet another fake-out with more major spikes coming! We’ve been burned at least twice so far in this interminable wave.

FYI: If someone caught COVID at Thanksgiving dinner on Thursday, they’d have developed symptoms circa yesterday. If they went and got tested today, that positive result would show up in our data around Friday of this week.

A second-order bump — people infected by those who got infected on Thanksgiving — would take another week or so to appear in the data.

Not saying this WILL happen in meaningful numbers, just when to look for it if it does.

SOME people will have caught #COVID19 at Thanksgiving dinners who would not have ordinarily been exposed. That’s not controversial.

The question is *how many* — whether it is enough to show up in our statewide data, to overwhelm other factors affecting disease spread.