1/28 Here's a thread, in which I prove @ChrisJLarson @SenChrisLarson correct about the damage done to Wisconsin by holding an election in the middle of a pandemic, and offer some hard data that puncture theories that Wisconsin is actually Incredibly Safe™.
2/ Within the last couple of weeks, @DHSWI has been releasing #Wisconsin #COVID19 data based on date of symptom onset or diagnosis. This is important, because the day that new cases are announced is never the day that new cases are diagnosed. dhs.wisconsin.gov/covid-19/count…
3/ I've got two charts for you here, this first one comparing the number of cases announced (blue) with the symptom-onset cases (red) by day, through 5/27, the last non-"preliminary" day in the data. There is clearly a lag between onset and reporting. 
4/ The second chart compares the cumulative number of cases based on when they were announced (blue) and when there was symptom onset (red). Again, you can see the mismatch.
5/ Knowing the date of symptom onset helps officials better identify when, where, and possibly even how outbreaks are occurring. It's why if you test positive for the coronavirus, you should really talk to the contact tracers trying to get your info.
6/ There is a militant subset of some in this state, mostly conservatives, who reject the idea of contact tracing and, in some cases, testing. "This is just to increase state control!" they say.
7/ The Venn diagram of those who oppose contact tracing and who insist the virus is overhyped is, as one would expect, a circle. They are boosted by two groups: Wisconsin Republicans and Wisconsin media.
8/ The media declared that the April 7 election had no effect on COVID19 cases in the state. There was no spike, they confidently reported, and @PolitiFactWisc even raked @ChrisJLarson over the coals for saying there was one. politifact.com/factchecks/202…
9/ The media also do not seem to understand patterns in Wisconsin data reporting, resulting in stories like this one that make the rounds as evidence that COVID19 is nothing to worry about. wbay.com/content/news/W…
11/ Of course, they had no counter-plan. Audio from the day-after meeting between @RepVos (who has blocked me here on twitter), @SenFitzgerald, and @GovEvers shows the GOP offers no suggestions for how to keep the state safe. jsonline.com/story/news/pol…
12/ Vos, in particular, was basically all ¯\_(ツ)_/¯ essentially wanting to throw all responsibility for dealing with the pandemic to local governments rather than allowing a state-led response. Fitzgerald was all, "We're open. It's done."
13/ So here's the big reveal: Below, you will see an annotated version of the first graph from earlier in the thread, with three key dates noted: when "Safer at Home" took effect*, the April 7 election, and the date of the court decision.
13.5/ *It took effect March 25, but there was an earlier order closing schools, for example, and many people, businesses, and municipalities took steps to limit exposure a couple of weeks earlier.
14/ I've helpfully added arrows. You'll see that once "Safer at Home" was in place, new cases per day plateaued based on symptom onset (red), even though it didn't look that way at the time as the state announced new cases (blue).
15/ For example, on 3/24, the state reported 41 cases, but 123 people actually had symptom onset that day. Flash forward to 4/7, the day of the election, there were 138 new cases announced (+97) and 133 new case onsets (only +10).
16/ After the election, check the arrow showing that about a week after, case onsets (red) spiked. I am not an epidemiologist so I am not sure whether they have a technical definition of "spike," but that's sure what it looks like to me.
17/ Median time from exposure to symptoms for this virus has been identified as about 5 days, so the fact that symptom onsets didn't spike on 4/8 is not a surprise. But on 4/13, there were 226 new case onsets, 93 more than 4/7, or a 69% jump.
18/ On the day of the court decision on May 13, 485 people had symptom onset, or a 264% jump over April 7. (FYI, the graph uses a 14-day rolling average for both announced and onset cases to smooth over day-to-day noise.)
19/ How can *anyone* look at the data and *not* come to the conclusion that the election had a devastating affect on infections here?
20/ The same is tru about the court decision. In the week following the end of "Safer at Home," the average number of reported cases fell. This certainly didn't help anyone out there understand how dangerous the "We're open. It's done" position is.
21/ But symptom onset was proceeding apace and, on 5/18 there was the single largest number of symptom onsets reported to date: 586. Most of those were probably not the idiots who packed the bars in celebration, but some probably were.
22/ DHS has been auditing/adjusting even the non-"preliminary" data, so I expect changes to the numbers leading into Memorial Day weekend, and of course we're still a week or so away from where we'd see effects from that.
23/ But here's an annotated version of the other chart, showing cumulative cases, with a note about exactly how big the difference is between announced cases (blue) and date-of-onset cases (red) on the 27th of the month.
24/ So based on the most complete set of data available, there was a gap of 2,637 cases on the last non-"preliminary" day between what the public new about and what was actually happening, a difference of 14%.
25/ The percentage difference has been falling as numbers get larger (math!), but the actual difference keeps getting bigger. This is not comforting. There could be up to 3,000 more active cases right now than we know about--an increase of 50%.
26/ The moral? State Republicans, by forcing an election upon us in a time of pandemic, absolutely created conditions for this disease to spike. By ending "Safer at Home" through a lawsuit and without replacement, has kept infections spreading.
27/ The media complicity in refusing to hold the GOP accountable and in enabling deniers to feel confident that nothing bad is happening has been near journalistic malpractice.
28/28 And Wisconsin citizens are paying the deadly price.
@threadreaderapp unroll
