Thread by @TheLawyerCraig

For those who follow my daily #Covid19 data updates, it's time for another period of likely volatility based on holiday reporting. Based on past holidays, here's what should happen:

(Thread)

Starting as early as today, reported numbers for tests, cases, and deaths will likely be lower than they would otherwise have been without a holiday.

This does *not* necessarily mean that (1) they will drop, or (2) every state/region will see the same amount of volatility.

This artificial deflation will likely continue through ~Tuesday of next week.

Then, the opposite happens! For 3-4 days, states generally report their current "pipeline" numbers PLUS any backlog numbers. We should see that from ~Wed-Sat of next week.

But it doesn't end there. Because I report 7-day averages, those averages end up higher than they otherwise would have been for the 3-4 days following the backlog reporting. Why? Because when you have 3 days normal + backlog & 4 days normal, the totals (and averages) are higher.

So the entire volatility period generally doesn't end for ~2 weeks from when it started. Here's a cheat sheet:

4/2 - Begin artificial low reporting
4/7 - Begin artificial high
4/10 - Normal reporting
4/16 - Full week of normal (7-day average should even out by today)

Generally, the hospitalization and ICU numbers don't see holiday volatility to nearly the same extent. There is always a weekend dip (compared to trend), and that can get exacerbated a bit by the holidays, but usually sorts itself out much more quickly.

Just keep the above in mind before trumpeting to the world "We're heading back down!" on 4/6, or "OMG did you see the rising case count in [State]!?" on 4/8, or "Wow, that 7-day average is SCARY!" on 4/12.

Be an elevated data connoisseur instead...