Map courtesy of datawrapper.de. Data source for this and other graphs of new case counts: Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 8/11/2021, from https://github.com/nytimes/covid-19-data.” The NY Times U.S. tracking page may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html.
Florida error, explained
The numbers I posted yesterday were wrong, due to a mixup in reporting of the Florida data. Because Florida accounts for such a large fraction of new cases, that in turn threw off the U.S. trend number.
You can read news reporting on this issue, because more-or-less everybody's numbers were wrong yesterday, including CDC. None of the reporting explains the issue, and some appears purposefully biased.
From where I sit, the problem is that Florida released its data a day early, but did not clearly indicate that. I noted the oddity of Florida's release yesterday. The upshot is that everybody was supposed to have intuited that there needed to be an extra day of zero new cases added to Florida's counts, but not actually reported by Florida.
It's easier just to show it.
Yesterday, the "last seven days" of data from Florida looked like this, in terms of reported cases / 100K.
| 8/9/2021 | 279.376 | 0 | 0 | 230.306 | 99.318 | 83.542 | 251.573 |
Today, the last seven days of data look like this:
| 8/10/2021 | 279.262 | 0 | 0 | 0 | 230.306 | 99.318 | 83.542 |
Note the extra zero. Florida reported the equivalent of 279 cases/ 100K yesterday. But they actually meant zero for yesterday, and 279 cases / 100K for today.
Unfortunately, with that "fix", the data now look a little bit too good, as you can see in the chart of growth rates below. So I'm still far from sure that we're actually looking at the accurate data out of Florida. In fact, I'm frankly confused as to what, exactly, Florida is reporting.
We now return to our regularly scheduled programming.
The (correct) trend data through 8/10/2021
With the Florida data reporting error now corrected, things look a lot more stable. The U.S. is now at 36.5 new cases/ 100K / day. Louisiana remains the hardest-hit state, and Florida is second with just under 100 new cases / 100K / day.
New cases continue to increase almost everywhere. But as you can see from the map embedded above, there is still a large north-south divide in terms of the level of new cases, with the Gulf Coast being more-or-less the epicenter of the Delta wave in the U.S.
Most importantly, the growth rate for daily new cases is going down. The national trend is visibly starting to curve downward (when graphed in logs, so that the slope is the growth rate).
Unfortunately, if I graph that, there's a sudden kink in the curve, concident with this Florida data reporting issue. So I now wonder if there wasn't or isn't some ongoing screw-up in the Florida data. Anyway, FWIW, here's the seven-day growth rate, graphed.
By eye, the graph of the entire U.S. pandemic history appears little changed.
Maybe I can see the top of the hill starting to form in the graph of all states. Maybe not.
Florida and Louisiana remain above their prior peak in cases / 100K / day.
On the plus side, Missouri, where this outbreak originated in the U.S., now appears clearly to have peaked. That's about 9 weeks from start to peak. If the U.S. follows the same timing, we'll peak around the first week of September.
