Nate Silver

Nate Silver @NateSilver538

@CarlBialik @SethS_D As far as I can tell, there are at least three schools of thought on how R0 is used in practice, so I'm just saying R by which I mean "how fast it's spreading in practice". — PolitiTweet.org

Nate Silver @NateSilver538

Can you find some patterns here? Sure. Fortunately the states with the very worst outbreaks (NY, MI, LA) are getting better. Weather may be a factor. Density may matter. Some states that never issued stay-at-home orders are having issues. But it's weird & idiosyncratic data. — PolitiTweet.org

Nate Silver @NateSilver538

Finally, 8 states have already fallen more than *50%* from their peaks. AK HI ID LA MT VT WA WY — PolitiTweet.org

Nate Silver @NateSilver538

8 more states have fallen 25-50% from their peaks: FL ME MO NY OK OR TN SC — PolitiTweet.org

Nate Silver @NateSilver538

8 states have fallen by 10-25% from their peaks. This starts to count as real progress. These are: AL CO MI NV PA TX WV WI — PolitiTweet.org

Nate Silver @NateSilver538

5 more states have peaks that occurred a week or more ago, but have fallen by less than 10% from the peak. These are: GA MD NM NJ UT* * increased testing volume likely a factor in UT So basically, half the states are still at or near their peaks in reported cases. — PolitiTweet.org

Nate Silver @NateSilver538

Another 6 states are not technically at their peaks, but the peak has come within the past week so it's hard to say it's over yet. These are: CT * KS KY MA * NC SD * increased testing volume likely a factor in CT & MA — PolitiTweet.org

Nate Silver @NateSilver538

While you might look at the national data and think things are going all right, 15 states + DC are *currently* at their peaks. These are: AR AZ CA & DC DE IA IL IN MN MS ND NE NH OH RI * VA & testing data is murky in CA * increased testing volume likely a factor in RI — PolitiTweet.org

Nate Silver @NateSilver538

Note: this is not ideal, because it doesn't account for changes in a state's volume of testing. But since testing has been quite stagnant in most states—not a good thing, BTW!—that probably matters less than usual. Where testing is a big factor, I will note that. — PolitiTweet.org

Nate Silver @NateSilver538

It really is under-appreciated how different the COVID-19 trajectories look in different states. Here's a quick THREAD with another simple way to summarize the data. Take a 7-day average of a state's newly-reported cases. Where does it compare to a state's peak? — PolitiTweet.org

Nate Silver @NateSilver538

@dansinker At least 5 or 6 of them are nice so don't sweat the other 3,299,994. — PolitiTweet.org

Nate Silver @NateSilver538

A slow but steady increase in "birthday party" searches. No change yet for "dinner party" though. https://t.co/7xkv1jmFX3 — PolitiTweet.org

Nate Silver @NateSilver538

I've been tracking Google searches for various kinds of party-related terms as a sign of lockdown fatigue... — PolitiTweet.org

Nate Silver @NateSilver538

There's this notion that people without domain expertise produce a lot of crap but the occasional brilliant finding. But on coronavirus it seems the opposite. As far as I can tell, there's lot of fairly solid B+ stuff. But the occasional disastrously wrong F- study. — PolitiTweet.org

Nate Silver @NateSilver538

Or there may be a lot of places where R is <1 among the general population, but R>>1 for critical workers and among other populations. A lot of the bad outbreaks recently are in places like prisons and meat-packing plants, not the general population. — PolitiTweet.org

Nate Silver @NateSilver538

One question we might ask from this is whether US-style social distancing is truly enough to get R<1 on its own, or instead it needs some "help" from herd immunity, warm weather, or — in some of these low-population states —geographic isolation + contact-tracing. — PolitiTweet.org

Nate Silver @NateSilver538

Generally the states doing better either: 1) are quite isolated (e.g. HI) 2) had very bad outbreaks already (e.g. MI), or 3) are in the Deep South, which may (?) reflect hotter weather If a state doesn't fall into one of those categories it may not be making much progress. — PolitiTweet.org

Nate Silver @NateSilver538

The list of *declining* states is harder because some states on the list haven't been great about reporting test volume. But some places that semi-robustly seem to be doing better include NY, VT, AK, HI, MI, MT. Probably LA too but their data on negative tests has been erratic. — PolitiTweet.org

Nate Silver @NateSilver538

Here are states with big recent increases, as measured by new cases recorded in the previous 7 days as compared to the 7 days prior: ND +191% OH +157% IA +105% NE +83% NH +54% DE +40% AR +40% MN +39% MS +39% CA* +33% * A bit misleading—CA level after accounting for test volume. — PolitiTweet.org

Nate Silver @NateSilver538

New York's numbers remain quite encouraging, as I've tweeted about, but that's a good news/bad news situation when looking at the national numbers. It means the rest of the country isn't making that much progress since most of the progress we are seeing comes from NY (+ NJ/CT). — PolitiTweet.org

Nate Silver @NateSilver538

New York's numbers remain quite encouraging, as I've tweeted about, but that's a good news/bad news situation. It means the rest of the country isn't making that much progress since most of the progress we are seeing comes from NY (+ NY/CT). — PolitiTweet.org

Nate Silver @NateSilver538

Tuesdays are often bad days because of how the data is reported. This was a Tuesday and it was a real bad day. Big one-day spikes in deaths reported in MA, PA, NJ and MI, likely reflecting reporting lags or efforts to count deaths that were missed before. — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking. Newly-reported deaths: Today: 2,674 (new high) Yesterday: 1,528 One week ago (4/14): 2,299 Newly-reported cases: T: 27K Y: 23K 4/14: 26K Newly-reported tests: T: 152K Y: 138K 4/14: 147K Share of tests positive: T: 18% Y: 17% 4/14: 18% — PolitiTweet.org

Nate Silver @NateSilver538

Keep in mind the context here. Do NYC and Lombardy possibly have meaningful levels of herd immunity? Possibly (enough that it may show up in the data) but exactly how much is hard to say (probably far from complete). Think about how bad things had to be to get there, though. — PolitiTweet.org

Nate Silver @NateSilver538

Positive test rate by week in Lombardy, Italy Week ended 3/24: 47.9% 3/31: 33.0% 4/7: 20.4% 4/14: 16.2% 4/21: 10.6% Positive test rate by week in rest of Italy Week ended 3/24: 17.5% 3/31: 16.0% 4/7: 14.0% 4/14: 11.8% 4/21: 9.9% — PolitiTweet.org

Nate Silver @NateSilver538

Also relevant here is Italy, where there’s been a much steeper decline in positive test rates in Lombardy, the worst-affected region, than in the rest of the country. Both peaked during the week of 3/24 but Lombardy has fallen notably faster: — PolitiTweet.org

Nate Silver @NateSilver538

Yeah, it is not crazy to ask questions about this stuff, particularly given how we're now seeing notably steeper deceleration in NYC than upstate. — PolitiTweet.org

Nate Silver @NateSilver538

@PSU_Ithaca Testing has been decreasing a bit over the past week. I think this may be a case where there is less demand for tests in hospitals, but testing isn't yet ramping up in other settings. — PolitiTweet.org

Nate Silver @NateSilver538

New York City had *seemed* have an R of around 0.9, with a fairly slow decline from a plateau. But the decline has accelerated in the past week or so and the past few days of testing data suddenly makes it look more like an R of 0.7. More like 0.8 in the suburbs and 0.85 upstate. — PolitiTweet.org

Nate Silver @NateSilver538

The positive test rate continues to plummet in New York City, to 28.5% in tests disclosed today. It was 42.5% a week ago and peaked at 59.4% on March 29. https://t.co/1w2VrzC4FG — PolitiTweet.org