Nate Silver

Nate Silver @NateSilver538

RT @WesleyLowery: As always, everything can be true. The violence can be the result of black residents, and also white anarchists and anti… — PolitiTweet.org

Nate Silver @NateSilver538

@SeanTrende Yeah, I think work from home is a big part of this. I'd guess that 95% of my commutes to work are via subway but it's 80% bikes/cabs/walks for other trips. Obviously that's n=1 but my general sense is that NYC is VERY dependent on transit for work, a bit less so for other trips. — PolitiTweet.org

Nate Silver @NateSilver538

@SeanTrende I mean I think it shows that people are making fairly smart choices. Friends I know in NYC who seem increasingly fatigued with social distancing still have no desire to take the subway unless they absolutely have to. — PolitiTweet.org

Nate Silver @NateSilver538

@SeanTrende I mean I think it shows people may be making fairly smart choices. — PolitiTweet.org

Nate Silver @NateSilver538

FWIW: Apple's mobility trends data shows a very large increase in mobility (both driving and walking) in Minneapolis, but not really in other cities where there's protests and/or civil unrest. https://t.co/32iSFYgSGd https://t.co/QtVoRC24gt — PolitiTweet.org

Nate Silver @NateSilver538

A lot of models try to infer the # of cases from the # of deaths and I wonder what this does to those. If we've gone from capturing only say 30% of actual COVID deaths in mid-March to 60-70% at the peak to 80-90% now, that could mess with your estimates of R and other parameters. — PolitiTweet.org

Nate Silver @NateSilver538

So, the actual peak of COVID-19 deaths may have been a bit earlier, higher and steeper than the official tally shows. A lot of deaths that happened early (before adequate testing) were missed. We're missing fewer deaths now that testing has improved. — PolitiTweet.org

Nate Silver @NateSilver538

RT @sfrostenson: "Law and order was about the restoration of a certain social configuration favorable to white Americans as much as it was… — PolitiTweet.org

Nate Silver @NateSilver538

At the same time... it's important to look at the data in the aggregate. It's a big country. You can ALWAYS find SOME place where cases are increasing. Some of those increases will be real. Others may reflect increased testing or some quirk or change in how the data is reported. — PolitiTweet.org

Nate Silver @NateSilver538

@conorsen I don't know what to expect at this point. But, yeah, it's plausible that cases will rise but deaths will not rise as much. Especially if we can do a better job of protecting nursing homes and/or some nursing homes had it badly enough already to reduce susceptibility. — PolitiTweet.org

Nate Silver @NateSilver538

I cautioned people when states to began opening up that the changes were fairly incremental; hence the effect on the data would likely be fairly marginal, also. But now you have far more people out and about, states in deeper stages of re-opening, some big cities opening up, etc. — PolitiTweet.org

Nate Silver @NateSilver538

Remember, though, that these numbers are backward-looking. There has probably been enough time to say that the initial, slow, cautious re-openings that began in late April aren't likely to cause huge spikes. However, there has been much more relaxation of distancing since then. — PolitiTweet.org

Nate Silver @NateSilver538

The story continues to be one of slow, steady progress. The 7-day average death total has fallen for 21 days in a row. And cases have ticked downward a bit after plateauing last week, despite increased testing...in fact, there were a record number of tests today, almost 500K. — PolitiTweet.org

Nate Silver @NateSilver538

US numbers via @COVID19Tracking: Newly reported deaths Today: 1,186 Yesterday: 1,353 One week ago (5/22): 1,280 Newly reported cases T: 23K Y: 23K 5/22: 24K Newly reported tests T: 493K Y: 415K 5/22: 394K Positive test rate T: 5% Y: 6% 5/22: 6% — PolitiTweet.org

Nate Silver @NateSilver538

RT @ABC: The first dog in the U.S. presumed to have COVID-19 may not have been infected with the virus after all, according to test results… — PolitiTweet.org

Nate Silver @NateSilver538

@jodyavirgan Not exaaaaactly confidence inspiring, one might say. — PolitiTweet.org

Nate Silver @NateSilver538

I am hoping the bike/transit people in my feed are wrong about how NYC has used its time to prepare (they are not pleased) but usually they are right about stuff. — PolitiTweet.org

Nate Silver @NateSilver538

Here we go. The metrics in NYC have been good lately, but obviously a lot of uncertainty as to how the reopening will go. Note that it's in several stages; not a lot of highly visible stuff in Stage 1. — PolitiTweet.org

Nate Silver @NateSilver538

On the contrary, the New York Times's analysis would seem to suggest that Florida, Georgia and Texas are in the middle of the pack. They don't have particularly high excess mortality, although almost surely they have some uncounted deaths. https://t.co/iTbN9SNc1d https://t.co/jziNhBzgbJ — PolitiTweet.org

Nate Silver @NateSilver538

It's well established that COVID-19 death tolls are undercounts. That's important for people to understand, and it's important for people to look at excess mortality data. There's no particular evidence that there's *more* undercounting in Florida, Georgia or Texas. — PolitiTweet.org

Nate Silver @NateSilver538

To be honest, it's a bit of a red flag when people are use (highly politicized) FL, GA or TX to illustrate (often dubious) points about coronavirus data. Many other states opened on roughly the same timelines so better to look at the data in a more comprehensive way. — PolitiTweet.org

Nate Silver @NateSilver538

@JamesSurowiecki @mattyglesias @foxjust Yeah, and in theory the underreporting issues should be getting better over… https://t.co/GGiMqCbrhn — PolitiTweet.org

Nate Silver @NateSilver538

@JamesSurowiecki @mattyglesias @foxjust But there's not too much reason to think this is different in Florida than… https://t.co/DVCtGJm4pp — PolitiTweet.org

Nate Silver @NateSilver538

@JamesSurowiecki @mattyglesias @foxjust Right, and it's important to point out that official death count estimates are generally too low and the actual death tolls are probably something like 1.2x to 1.5x the reported figures depending on when and where you're looking at. — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias @foxjust I just think people need to self-monitor a lot better in the category of "quick-and-dirty interpretations of the data that doesn't match the consensus of the evidence but DOES match my priors". Like, those interpretations are going to be wrong some hugely high % of the time. — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias It stems from people not having their priors confirmed and grasping at straws. — PolitiTweet.org

Nate Silver @NateSilver538

There hasn't been. People are: 1) mixing and matching two non-comparable data sources. https://t.co/l3mHb6o2c4 2) not accounting for the fact that deaths can be classified as pneumonia *and* COVID-19. https://t.co/EFBMaB1oJP — PolitiTweet.org

Nate Silver @NateSilver538

Here's what it looks like if you take the 7-day rolling average. The numbers have been relatively flat in Wisconsin, not down, which isn't great, and in general I worry about the Upper Midwest a bit. But the implication of a major spike or surge is pretty misleading. https://t.co/0HPZNKbaAa — PolitiTweet.org

Nate Silver @NateSilver538

Way too many misleading stories like these based on one day's worth of figures. Wisconsin reported 22 deaths on Wednesday but that was after three very slow days (3, 4, 3) Sunday thru Tuesday, which likely reflected lags because of the 3-day weekend. https://t.co/CNRc9rlmQp — PolitiTweet.org

Nate Silver @NateSilver538

Despite an utter lack of leadership from the federal government, many states are implementing the steps experts advised (much more testing, contact tracing programs, encouraging mask-wearing, still banning mass gatherings) and it's sort of weird they don't get more credit for it. — PolitiTweet.org