I’ve been following the US case trackers pretty closely, and have been a little surprised at the speed with which the US curve has flattened since the reporting protocols were changed to bypass the CDC. Florida in particular has turned right around.
Aside from waiting a few weeks to see if Florida turns into a nice dubious bell curve* or rolls off gently like all other places have done, does anyone have any insight into whether this flattening is real?
one of my favorite apocryphal statistical stories is the origin of Benford’s law arising from a close examination of accounting records, where too many leading 9s revealed a crook.*
**Not to imply in any way that we’re dealing with crooks in this case.
What? The US recorded the most cases ever three days ago. Are you talking about cases going down for two days and asking if that’s statistically meaningful?
As of today their running averages have been flattening out for 5 or 6 days, now — long enough to see a qualitative change from what had appeared to be pretty strong growth. Is there a better data source?
The 7 day moving average had been climbing an average of 1500 cases per day for just over a month up until that day you mentioned as the most cases ever. In the five days since, it’s risen less than 1000 total.
I don’t think 5 days is enough to declare a trend, but it is a little suspicious that it happened at the same time as the new reporting protocols. We’ll have to wait and see.