COVID-19 2nd wave - Texas only - Stats and such

[BabaYaga 15305]

4 minutes ago, drt said:

God fucking damnit this is the texas thread.  Take this shit to the general thread or the CR thread.

LOL, you are 100% right.  But people need to take this time to take care of themselves.  

[Anastasis 23481]

What is interesting to me is to compare the timeline of re-opening in CA and TX (Bars in CA for example only re-opened in Mid June, Bars in TX in mid May; CA restaurants in mid June, Tx in early May).

That lag time in reopening is pretty critical.  But when we look at daily new cases, CA and TX both appear to have broke into a new slope ~ June 15. Almost identical timing, despite substantial differences in re-opening timelines. 

 

Edited July 2, 2020 by Anastasis

[Anastasis 23481]

8 minutes ago, Brisketexan said:

Cool.....now do New York, which had MASSIVE protests.....

New York is not a good control group for anyone in this discussion, except maybe Italy. 

[BabaYaga 15305]

1 minute ago, Anastasis said:

What is interesting to me is to compare the timeline of re-opening in CA and TX (Bars in CA for example only re-opened in Mid June, Bars in TX in mid May; CA restaurants in mid June, Tx in early May).

That lag time in reopening is pretty critical.  But when we look at daily new cases, CA and TX both appear to have broke into a new slope ~ June 15. Almost identical timing, despite substantial re-opening timelines. 

 

"Re-opened" has to be quantified.  It's not binary:  open or close.  Based upon the phases, and percentages of patrons allowed.  

[smoky 840]

What is interesting to me is to compare the timeline of re-opening in CA and TX (Bars in CA for example only re-opened in Mid June, Bars in TX in mid May; CA restaurants in mid June, Tx in early May).
That lag time in reopening is pretty critical.  But when we look at daily new cases, CA and TX both appear to have broke into a new slope ~ June 15. Almost identical timing, despite substantial differences in re-opening timelines. 


 

What's Georgia look like compared to Texas and California?

Didn't they have a similar reopening to us?

[Huckleberry 33073]

4 minutes ago, BabaYaga said:

"Re-opened" has to be quantified.  It's not binary:  open or close.  Based upon the phases, and percentages of patrons allowed.  

That's pretty funny considering you were the one who made it binary by saying they never opened and were wrong. Good job googling after your lie was exposed, though.

[Anastasis 23481]

2 minutes ago, smoky said:

What's Georgia look like compared to Texas and California?

Didn't they have a similar reopening to us?

Pretty similar to TX and CA, new slope starts around June 15. 

[gmr548 16607]

What is interesting to me is to compare the timeline of re-opening in CA and TX (Bars in CA for example only re-opened in Mid June, Bars in TX in mid May; CA restaurants in mid June, Tx in early May).
That lag time in reopening is pretty critical.  But when we look at daily new cases, CA and TX both appear to have broke into a new slope ~ June 15. Almost identical timing, despite substantial differences in re-opening timelines. 


 

Just by eyeballing I think you could argue that because CA's daily new cases never totally flattened (looking at you LA), just kept rising at a shower pace, the time to going exponential would be quicker because they started from a higher baseline of community spread.

That take could of course be totally off and should be shredded accordingly if it is.

[MNLonghornFUKM 8817]

Cool.....now do New York, which had MASSIVE protests.....https://www1.nyc.gov/site/doh/covid/covid-19-data.page   That's just one example.
And, to extrapolate your conclusion correctly, you believe that large groups, outdoors, even when many are wearing masks, are not just a significant source of transmission, but are among the BIGGEST sources of transmission?  Because that's what your conclusion means.  Is that your conclusion?  Pay attention to the implication of that conclusion, and what it means for beaches, parks, sporting events, and every other activity where there might be a group outdoors.
 

NY was at 20% immunity in late April, which by now is even higher. There’s so many carriers there that the virus isn’t going to bounce around as easily than places that weren’t hit as hard

[Brisketexan 105275]

8 minutes ago, MNLonghornFUKM said:

NY was at 20% immunity in late April, which by now is even higher. There’s so many carriers there that the virus isn’t going to bounce around as easily than places that weren’t hit as hard

Cool.  So, we should mark you down in the column of "outdoor activities involving groups are too high risk" then?  Just want to know where you stand on closing potentially crowded beaches, parks, pools, etc.

[Hpara759 1780]

45 minutes ago, BrazilHorn said:

What I can say in driving through West Texas recently is that NO ONE was wearing a mask. (minus the workers and bulk of them had masks on chins with noses sticking out)

It is like towns like Lubbock. Amarillo,  etc just think that Covid is made up

I think Lubbock, Randall and Potter counties know that it's real......2,457, 899, 2,913 cases respectfully

[NoName 16369]

please...

[JBJ 3664]

1 hour ago, Anastasis said:

Can you break down the methodology that you use to arrive at these conclusions a bit further?

Sure.

The important model assumptions are (these were built around the US data) (1) results get posted an average of about 11 days after exposure (this is a distribution - 3sig is 5-17 days) (2) there's roughly 10x the exposures as cases reported.

----

I've created 4 windows for when people were exposed:

Phase 2 transition: May 18-May 27

Phase 2 without protest: May 27- June 5

Protest: June 5 - June 15

Post-protest /pre-knowledge of surge: June 15 forward

----

Method:

1) Use the model to get the transmissivity for each window.  Compare overall transmissivity, 

2) Make sure the overall is what is in the daily numbers as well. Transmissivity should clearly jump from one rate to another in the definable periods despite daily noise.

3) Increases in transmissivity associated with the protest period that don't exist in the post-protest period are likely due the protests themselves.

4) Disclaimer: Data isn't complete yet for post-protest period, but because jumps do show up per #2, I'm pretty confident in this.

Edited July 2, 2020 by JBJ

[Guest Lobo]

"I'm not racist but black people protesting in large groups outdoors spread Coronavirus, but white people in partying in large groups outdoors didn't spread Coronavirus."

I mean, look down on other races to make yourself feel superior.  Or deny science.  But it's just really not a good year to be a science-denying racist.  Seriously, we just don't have room for that in 2020.  You can be one shitty thing this year, we're all doing that.  But you gotta pick a lane.  We all gotta focus in 2020.  We can be one shitty thing, but otherwise we gotta buckle up and do our part for our neighbors, for our nation; for our relatives, for our republic. 

Edited July 2, 2020 by Lobo

[BrazilHorn 6630]

12 minutes ago, Hpara759 said:

I think Lubbock, Randall and Potter counties know that it's real......2,457, 899, 2,913 cases respectfully

Maybe,  but not real enough to wear a mask right. (or at all)

Was in gas station in Snyder and Zero people outside of me had a mask on. Workers all had theirs on wrong.

Funny this is saw exact same thing in Pueblo Colorado.

May just be a small town belief that it won't happen to them.

 

[Huckleberry 33073]

15 minutes ago, Hpara759 said:

I think Lubbock, Randall and Potter counties know that it's real......2,457; 899; 2,913 cases respectfully

Fixed so it doesn't look like those three counties might have over 24 billion cases.

[BradInATX 9208]

58 minutes ago, MNLonghornFUKM said:

NY was at 20% immunity in late April, which by now is even higher. There’s so many carriers there that the virus isn’t going to bounce around as easily than places that weren’t hit as hard

You mean like... Minnesota?

 

 

Edited July 2, 2020 by BradInATX

[Reynolds Woodcock 4799]

44 minutes ago, MNLonghornFUKM said:

NY was at 20% immunity in late April, which by now is even higher. There’s so many carriers there that the virus isn’t going to bounce around as easily than places that weren’t hit as hard

While it is true that NYC has an overall ~20% positive antibody test rate, unsurprisingly that 20% is not uniformly spread. The same study shows that white people have an ~6% positive rate (Latinos, on the other hand, have an ~35% positive rate).  Studies following the huge early protests estimate that 60-75% of the protesters were white. The largest protests also took place in the areas of the city that weren't hit as hard (lower Manhattan and west Brooklyn).

      

[BradInATX 9208]

50 minutes ago, JBJ said:

Sure.

The important model assumptions are (these were built around the US data) (1) results get posted an average of about 11 days after exposure (this is a distribution - 3sig is 5-17 days) (2) there's roughly 10x the exposures as cases reported.

----

I've created 4 windows for when people were exposed:

Phase 2 transition: May 18-May 27

Phase 2 without protest: May 27- June 5

Protest: June 5 - June 15

Post-protest /pre-knowledge of surge: June 15 forward

----

Method:

1) Use the model to get the transmissivity for each window.  Compare overall transmissivity, 

2) Make sure the overall is what is in the daily numbers as well. Transmissivity should clearly jump from one rate to another in the definable periods despite daily noise.

3) Increases in transmissivity associated with the protest period that don't exist in the post-protest period are likely due the protests themselves.

4) Disclaimer: Data isn't complete yet for post-protest period, but because jumps do show up per #2, I'm pretty confident in this.

 

- No controls
- Terrible assumptions ("Transmissivity should clearly jump from one rate to another in the definable periods despite daily noise." - nope)
- Doesn't use any data around actual increases in socialization/mobility, just "I bet people started moving around in these time periods"
- No analysis of whether you actually have a correlation across different data points
- No covariate to deal with places that did and did not have protests

 

I took all your research and made you a graph that represents your findings, if you want some visuals to add on to your Nobel prize application

 

 

If you want to see what the analysis you attempted looks like with people that actually know what they're doing, try this:

https://www.nber.org/papers/w27408.pdf

 

 

Edited July 2, 2020 by BradInATX

[Doc Reeves 6434]

Welp. Had a false fever alarm with my daughter at preK today and went to pick her up. This place knows their shit and while I was speaking to the director she told me that they have just been briefed by the city (SA) that 15 daycares have reported child COVID cases. We have 3 birthdays coming up over the next week or so going back into quarantine again sucks. The wife and I are probably going to take them out until stuff settles down. Fortunately, we fully expected this scenario at this exact time because of, you know, High School heath class and have planned accordingly. 

Edited July 2, 2020 by Doc Reeves

[dcar00 14344]

why are we talking about cases? I thought this was about deaths.  how is Texas doing in deaths?

[Doc Reeves 6434]

2 minutes ago, dcar00 said:

why are we talking about cases? I thought this was about deaths.  how is Texas doing in deaths?

I know a Dr. dealing with this at the Methodist here in SA and he is absolutely adamant that dying isn’t the worst thing that can happen to you if you get this shit.

[Skipper 6193]

After being outdone by Tarrant County the last couple of days (with respect to cases at least), Dallas coming in hot with +708 and 7 deaths.  Your move Ft. Worth.

[Foosters 17883]

2 hours ago, BabaYaga said:

Then explain CA, which never opened, yet saw a SPIKE in cases.  

I can wait....

Never opened? I live in California. Shit's been open man. Here's the Gaslamp district in San Diego in early June. Bars open. 

https://www.latimes.com/california/story/2020-06-17/column-wonder-why-so-many-people-arent-wearing-masks-heres-how-they-explain-it

Edited July 2, 2020 by Foosters

[Anastasis 23481]

37 minutes ago, BradInATX said:

- No controls
- Terrible assumptions ("Transmissivity should clearly jump from one rate to another in the definable periods despite daily noise." - nope)
- Doesn't use any data around actual increases in socialization/mobility, just "I bet people started moving around in these time periods"
- No analysis of whether you actually have a correlation across different data points
- No covariate to deal with places that did and did not have protests

Some of this is reasonable critique, but not uniformly. 

If I am understanding what he is doing, he is essentially doing a form of time series analysis.  External controls are not necessary, as it is essentially self-controlled. In certain cases this may be preferred over other approaches because it allows you to essentially remove the influence of potentially confounding factors that are present when using external controls (for example, NY as a control group).  It's not clear to me at what level he is analyzing (Travis County, Texas, or US).  I presume TC based on the recent discussion here, but that might be wrong.  As I have argued previously, I think that the impacts are regional.  The drivers of what is playing out in Austin are different than what is playing out in Corpus. If you run that type of analysis using TC data, I think that you will find a different answer than if you ran it with Nueces county data. 

It is not clear to me why the transmissivity assumption is terrible.  "Nope" doesn't really strike me as a critique worthy of the effort JBJ's been putting into his modeling. 

I am not sure why the third point is relevant.

Covariate for protests is not really relevant.  The protest influence is captured by the time variable. 

 

I may be misinterpreting his analysis, and if so much or all of the above may be totally wrong. 

Edited July 2, 2020 by Anastasis

[elfenix 34331]

32 minutes ago, dcar00 said:

why are we talking about cases? I thought this was about deaths.  how is Texas doing in deaths?

why would this only be about deaths?

[BradInATX 9208]

13 minutes ago, Anastasis said:

Some of this is reasonable critique, but not uniformly. 

If I am understanding what he is doing, he is essentially doing a form time series analysis.  External controls are not necessarily, as it is essentially self-controlled. In certain cases this may be preferred over other approaches because it allows for a relatively good level of many of the potentially confounding factors that are present when using external controls (for example, NY as a control group).  It's not clear to me at what level he is analyzing (Travis County, Texas, or US).  I presume TC based on the recent discussion here, but that might be wrong.  As I have argued previously, I think that the impacts are regional.  The drivers of what is playing out in Austin are different than what is playing out in Corpus. If you run that type of analysis using TC data, I think that you will find a different answer than if you ran it with Nueces county data. 

It is not clear to me why the transmissivity assumption is terrible.  "Nope" doesn't really strike me as a critique worthy of the effort JBJ's been putting into his modeling. 

I am not sure why the third point is relevant.

Covariate for protests is not really relevant.  The protest influence is captured by the time variable. 

 

I may be misinterpreting his analysis, and if so much or all of the above may be totally wrong. 

- A time series analysis is not even remotely appropriate here. A time series analysis might be effective if we were trying to figure out how long it takes to grow daisies in a temperature controlled room with all other variables being equal. It's incredibly simplistic and immature for what he's trying to do. I think you think you're being fancy by using that word but a time series analysis is literally what my MS Paint represents. 

- The impacts being regional doesn't have anything to do with anything. If you want to establish a correlation between two things, you need to do lots of things that I'm not going to go into on this thread. But looking at line chart with one line and about 60 data points is nowhere even close to that. If he's just looking at Travis County and saying "protests happen, cases go up" then that's literally like saying "I flipped a coin once and it was heads, so when I flip a coin it is always going to be heads" and if he did it on national data it's the same thing but with a really big coin. 

- The transmissivity assumption is terrible because he's using the outcome to define the inputs, which is bush league.

 

Frankie the weather man puts effort into his work too. At least he's not pushing an agenda.

Edited July 2, 2020 by BradInATX

[troph 24905]

So The NY Times has been a pretty good source of info. I like this page a lot https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html for its overall view of the country and the state breakdowns. The graphics are good too if you have anyone you are sending info to that might have trouble just looking at numbers.
So the other day I stumbled onto 2 really good articles that examine the probable “real” death toll from covid. https://www.nytimes.com/interactive/2020/05/05/us/coronavirus-death-toll-us.html focuses on the Unites States and breaks it down by state. https://www.nytimes.com/interactive/2020/04/21/world/coronavirus-missing-deaths.html attempts to tackle the question on a global level.
It’s some really interesting stuff if you are bored and have a little free time.

It’s going to prove out at 4-6x as deadly as the seasonal flu (instead of 20x) and more contagious with no knowledge of long term ramifications. Folks will claim it’s just like the flu but that doesn’t sound like the flu to me.

[Guest Lobo]

When you look at the spike map, it just doesn't scream "Black Lives Matter Protest Hotspots."  Yeah, some...but this 60% correlation claim is absolutely fucking insane.  

We're a football message board with a drinking problem and no sports to wax poetic about that suddenly had to pivot to a place of science and co-venting to one another during a lockdown.  We're entitled to throw up dumb shit, we ain't peer-reviewed.  But some of this lunacy probably begins to permeate behavior and lifestyle, which means you're going out in public with this stupid shit rocking around in your head.  And that means you're probably doing dumb shit when you're out in public, and that is going to effect me and my family.  That basic physics escapes the grasp of some of you is just delightful.  How you made it this far in life should be studied harder than the virus is being studied.

But yeah, it's just a bad flu season.  

Edited July 2, 2020 by Lobo

[Brisketexan 105275]

2 minutes ago, troph said:

It’s going to prove out at 4-6x as deadly as the seasonal flu (instead of 20x) and more contagious with no knowledge of long term ramifications. Folks will claim it’s just like the flu but that doesn’t sound like the flu to me.

This sort of reasoning, leaving out the second mathematical truth, is what blows me the fuck away.

If it is 4x-6x as deadly as the flu....AND it is more contagious (say, it's twice as contagious, for the sake of argument), then you're looking at 8x-12x the number of deaths as we see with the flu.  And that's a big fucking number.  Infection fatality rate is one relevant number.  The number of total infections is another relevant number.  Together, they tell us "how bad" this disease is.

All relevant data tells us that those numbers together yield a result of "it's pretty fucking bad."  But, we're an ignorant people, and math is for commies, so here we are.

[JBJ 3664]

 

21 minutes ago, Anastasis said:

Some of this is reasonable critique, but not uniformly. 

If I am understanding what he is doing, he is essentially doing a form of time series analysis.  External controls are not necessary, as it is essentially self-controlled. In certain cases this may be preferred over other approaches because it allows you to essentially remove the influence of potentially confounding factors that are present when using external controls (for example, NY as a control group).  It's not clear to me at what level he is analyzing (Travis County, Texas, or US).  I presume TC based on the recent discussion here, but that might be wrong.  As I have argued previously, I think that the impacts are regional.  The drivers of what is playing out in Austin are different than what is playing out in Corpus. If you run that type of analysis using TC data, I think that you will find a different answer than if you ran it with Nueces county data. 

It is not clear to me why the transmissivity assumption is terrible.  "Nope" doesn't really strike me as a critique worthy of the effort JBJ's been putting into his modeling. 

I am not sure why the third point is relevant.

Covariate for protests is not really relevant.  The protest influence is captured by the time variable. 

 

I may be misinterpreting his analysis, and if so much or all of the above may be totally wrong. 

It's Texas. 

Most if what he says is inapplicable.  It's a simple impact analysis.  The comparisons are the adjacent time periods.

Others are valid, but my post errs are on the side of the presumption.  I.E. Adjusting for increasing social interaction would make the protests look worse.  Same for geography, but I also don't think there are easily established borders.

The transmissivity jumps aren't an assumption, they are listed because they should be hypothetically observable.   And they are in fact observable.  If they weren't observable, I'd need to question my model and/or approach.  Transmissivity is akin to a measure of the curvature of whatever sinusoid path you are currently on.  If a discrete event impacts it, it's an immediate impact.   If the were no jumps, I'd be hard to point to any discrete event affecting transmission.

Edited July 2, 2020 by JBJ

[troph 24905]

I don’t know if math is for commies but it’s not for lawyers. That much I know.

 

Yeah it’s bad.

 

Though I’m starting to wonder if catching it is inevitable. /sad face

[Anastasis 23481]

11 minutes ago, BradInATX said:

- A time series analysis is not even remotely appropriate here. A time series analysis might be effective if we were trying to figure out how long it takes to grow daisies in a temperature controlled room with all other variables being equal. It's incredibly simplistic and immature for what he's trying to do. I think you think you're being fancy by using that word but a time series analysis is literally what my MS Paint represents. 

- The impacts being regional doesn't have anything to do with anything. If you want to establish a correlation between two things, you need to do lots of things that I'm not going to go into on this thread. But looking at line chart with one line and about 60 data points is nowhere even close to that. If he's just looking at Travis County and saying "protests happen, cases go up" then that's literally like saying "I flipped a coin once and it was heads, so when I flip a coin it is going to be heads" and if he did it on national data it's the same thing but with a really big coin. 

- The transmissivity assumption is terrible because he's using the outcome to define the inputs, which is bush league.

 

Frankie the weather man puts effort into his work too. At least he's not pushing an agenda.

You seem grumpy for the last workday of the week.  May I suggest you knock off and start happy hour early to get in the right frame of mind.

There is absolutely nothing wrong with an analysis based around comparing infection rates during discrete time periods.  It gets precisely to the crux of the issue. There are limitations sure, but any approach is going to have some limitations. For example that NBER paper you posted only rolls their data out to June 20, which may not be long enough given the lags between exposure and testing and reporting of results. They also do not account for the impacts of changes in non-protester behaviors during the protest periods.

We can all poke holes in each others data analysis, but at least JBJ is running some numbers and bringing some data to the discussion.  

[GreenspointTexas 9664]

Houston traffic is the worst ive seen in almost 4 months

 

 

gonna be a fun july

[pearlandhorn 3274]

3 hours ago, Hate said:

Are you feeling any better?

 

Edit: intended for pearlandhorn

Surprising woke up this morning with no fever and no muscle soreness. Just coughing a shit ton and have a lot of congestion.

[justhookit 4191]

1 minute ago, GreenspointTexas said:

Houston traffic is the worst ive seen in almost 4 months

 

 

gonna be a fun july

Almost like it’s the start of a holiday weekend or something.

[DDD Dad 14347]

4 hours ago, JBJ said:

Right now, it seems about 50%-60% of this surge is tied to protests.  That is the portion of transmission I am seeing already subside as of yesterday's numbers.

The other 40%-50% I attribute to various lack of social distancing / desire for normalcy.  It is hard to tie that directly to Memorial Day or Phase 2 because of the timing of these events vs the surge itself, but I'm sure all 3 (reopening, holiday, protests) play a role in the general behavior of people.

As stated before there was an increase associated with Phase 2, but it was a more subtle steady increase (R0~1.3) - not the surge we saw/are seeing (R0~2.2).

So you're just pulling percentages out of your ass?  

[dcar00 14344]

4 minutes ago, DDD Dad said:

So you're just pulling percentages out of your ass?  

as is everyone on this thread(and most of the scientists to be honest) when it comes to this virus  and mask wearing

Edited July 2, 2020 by dcar00

[Leeroy Jenkins 2412]

In my little red neck part of East Texas I was able to get a rapid test done yesterday within 10 mins of calling and got the results back 17 mins after they scratched the underside of my brain.  Negative

 

I found out I had been exposed (for a short time and outside) to a COVID positive person a few days before.    

[JBJ 3664]

9 minutes ago, DDD Dad said:

So you're just pulling percentages out of your ass?  

More or less.  I'll revisit and post the chart in 4-5 days.  The 10-day trend needs some time to catch up, and the data isn't all in yet.

Edited July 2, 2020 by JBJ

[Guest Lobo]

4 minutes ago, dcar00 said:

as is everyone on this thread(and most of the scientists to be honest) when it comes to this virus  and mask wearing

Mask wearing is a percentage pulled out of one's ass?  How's that?  

When fat people talk, spit particles come out of their exhausted fat mouths.  Those particles find their way onto other people's faces/airways.  that's how this is spreading in the south.  Because we are filled with fat, inconsiderate people.  Same way the flu gets passed around.  This thing just sucks more.

And JBJ---you're gonna post a 10-day trend in 5 days to prove a point about protests that basically ceased in most spiked counties 15 days ago?  I know science is hard for me.  And maths are hard for most.  But now you seem to be having difficulty with calendars?  

Edited July 2, 2020 by Lobo

[dcar00 14344]

6 minutes ago, Lobo said:

Mask wearing is a percentage pulled out of one's ass?  How's that?  

When fat people talk, spit particles come out of their exhausted fat mouths.  Those particles find their way onto other people's faces/airways.  that's how this is spreading in the south.  Because we are filled with fat, inconsiderate people.  Same way the flu gets passed around.  This thing just sucks more.

And JBJ---you're gonna post a 10-day trend in 5 days to prove a point about protests that basically ceased in most spiked counties 15 days ago?  I know science is hard for me.  And maths are hard for most.  But now you seem to be having difficulty with calendars?  

apparently you don't read so good but continue with the "aaaaaaaaargh, fat people, rubes, etc!" commentary.  I thought it was 9 days on a doorknob.  grab a beer, spit on the ground in your backyard, its Fourth of July Weekend!

[atomheartbevo 89778]

4 hours ago, JBJ said:

Right now, it seems about 50%-60% of this surge is tied to protests.  That is the portion of transmission I am seeing already subside as of yesterday's numbers.

The other 40%-50% I attribute to various lack of social distancing / desire for normalcy.  It is hard to tie that directly to Memorial Day or Phase 2 because of the timing of these events vs the surge itself, but I'm sure all 3 (reopening, holiday, protests) play a role in the general behavior of people.

Everything else has been dragged out already, but Forbes looked at credit card spending in the restaurant segment (which also includes bars), and it matched up with data from Johns Hopkins - there were clearly big jumps in areas where restaurants/bars/etc. began opening up and people could dine in.  And countering those areas, areas where grocery store spending was steady/increasing, saw lower covid levels.

We knew this was going to happen.   Our governor publicly said cases would rise as we started reopening.  This isn't rocket science.  Well, it kind of is in some ways, but the point is, places like Texas and Florida that reopened with few limitations (dropping mask requirements, etc.) saw a big spike.

The problem now is that we as a whole don't seem too intent on putting on the brakes, so we are going to drive over the cliff and see what happens.

 

 

[atomheartbevo 89778]

18 hours ago, South Austin said:

What things will be shutdown?

I've been told it's now July 6.  As for what, it could be masks in more areas, it could be making public gathering limits even lower, it could be gyms, could be retail.

But it won't happen before the 6th on a statewide basis.

[BradInATX 9208]

 

39 minutes ago, Anastasis said:

You seem grumpy for the last workday of the week.  May I suggest you knock off and start happy hour early to get in the right frame of mind.

There is absolutely nothing wrong with an analysis based around comparing infection rates during discrete time periods.  It gets precisely to the crux of the issue. There are limitations sure, but any approach is going to have some limitations. For example that NBER paper you posted only rolls their data out to June 20, which may not be long enough given the lags between exposure and testing and reporting of results. They also do not account for the impacts of changes in non-protester behaviors during the protest periods.

We can all poke holes in each others data analysis, but at least JBJ is running some numbers and bringing some data to the discussion.  

 

IMO bringing bad analysis to the table is worse than doing nothing at all. 

Since we're doing ridiculous analysis on single line graphs, I want to do one of my own, because it must be nice just to do whatever I want without having to use best practices or intellectual rigor. I do impact analyses and correlation analyses on hundreds of millions of dollars of revenue and spend on almost a daily basis for my clients, and it's a beating having to be, you know, scientific. So in the spirit of "Impact analyses", I wanted to do one of my own.

 

Germany started shutting things down early March. Cases still climbed. In fact, they didn't slow down until almost a full month later. So my "time series analysis" shows that the shutdown was ineffective.

but

On March 26th, this happened:

 

11 days later, cases started dropping dramatically.

So using a "time series analysis" I can reasonably conclude that Germany's shutdown efforts did not work, but that to beat Covid, we simply need to receive a 19-year old message in a bottle from Ireland.

I've put together another visual for us:

 

Science is fun

 

 

 

Edited July 2, 2020 by BradInATX

[atomheartbevo 89778]

I early voted, had my mask, used sanitizer, and got finger condoms for the touchscreen (and finger condoms can make a tiny dude's dick look huge and/or impress a date when he can't put it on).

But damn, the whole time I was in there, I was thinking "wow, these voting volunteers all have pre-existing conditions and/or are over the age of 65".

Hats off to those folks, because that took some guts on their part.

[JBJ 3664]

15 minutes ago, Lobo said:

Mask wearing is a percentage pulled out of one's ass?  How's that?  

When fat people talk, spit particles come out of their exhausted fat mouths.  Those particles find their way onto other people's faces/airways.  that's how this is spreading in the south.  Because we are filled with fat, inconsiderate people.  Same way the flu gets passed around.  This thing just sucks more.

And JBJ---you're gonna post a 10-day trend in 5 days to prove a point about protests that basically ceased in most spiked counties 15 days ago?  I know science is hard for me.  And maths are hard for most.  But now you seem to be having difficulty with calendars?  

Yeah.  The way data reporting works is that it lags. I need to be 17 days removed to get 99.7% of the cases reported from any single day of exposure.  It takes another 10-days for a 10-day trend to move where it wants to go.

Edited July 2, 2020 by JBJ

[MNLonghornFUKM 8817]

Cool.  So, we should mark you down in the column of "outdoor activities involving groups are too high risk" then?  Just want to know where you stand on closing potentially crowded beaches, parks, pools, etc.

What? I didn’t even come close to saying that. Outdoor is fine with me, I had my first beer league softball night on Monday.

[atomheartbevo 89778]

Alamo Drafthouse laying off more folks

https://www.kxan.com/news/coronavirus/alamo-drafthouse-to-lay-off-some-employees-because-of-covid-19-pandemic/

Pools, playgrounds, greenbelts, etc. closing in Austin for the weekend, but certain amenities will not be re-opening

https://www.kxan.com/news/coronavirus/austin-park-amenities-several-greenbelts-to-close-in-response-to-covid-19-spike/

Quote

Already, all PARD parks and facilities were scheduled to be closed from Friday July 3 through Sunday July 5. But this new memo explains that when Monday, July 6 comes around, the parks themselves will re-open but the park amenities will stay closed indefinitely, a spokesperson for PARD explained.

I have a lot to say on this.   Full-on basketball on the courts with shirts and skins, no masks, sweat and heavy breathing everywhere, bad and should be banned.  Tennis courts?  And are swimming pools hot-beds of covid?  Because it doesn't seem like it.  Golf courses?

BUT WAIT, IT GETS BETTER!

Quote

Amenities like playgrounds, volleyball courts, tennis courts and disc golf courses will be closed as part of this effort. The only amenities that will remain open are restrooms and water fountains, the city said. 

Facilities that belong to the parks department will be closed as well, including facilities that had reopened like cultural facilities, pools, golf courses, tennis facilities and park concessions. 

The city says that the Barton Creek and Bull Creek Greenbelts will also remain closed. The city’s in-person programming for the month of July will be suspended, including sports and fitness programming, summer camps and cultural programming.

So golf courses and tennis courts and some greenbelts are staying closed.

But fucking restrooms and water fountains are staying open.

[htown85 159]

 

Another drop from prior week.  7 day average dropping.  Encouraging.