Coronavirus pandemic

[Scott5114]

Why should anyone take a graph with no y-axis seriously?

Each one has a y-axis where the bottom is zero and the top is the high point of the line.  The purpose is to compare the shapes and directions, not the heights.  If each state had a y-axis as high as New York's, you could compare the heights, but it would make the changes in direction harder to see on the very short curves.

Except since they're not labeled, you have no frame of reference for any of the numbers. So you have no idea if the graph maker is doing stupid shit like this.

[Eth]

Generally you use running averages to smooth things out.  But this virus seems strange to me.  I'm not a statistician, so maybe there's a reason for this I'm not familiar with, but even on the national scale and global scale, large enough that you'd think the irregularities would smooth out, since not every place has the same irregularities, the daily new cases numbers follow a fairly regular wave pattern with peaks and troughs.  I would expect this would be more random, with an average daily number of X over a weeklong period, and each day staying mostly within a certain percentage of that number, but randomly either higher or lower.  In such a pattern, waves would sometimes happen by chance, but you wouldn't expect waves consistently.  It isn't a perfect wave pattern.  Sometimes it goes up, down, then up.  But it isn't nearly as random as I would expect.

On Worldometer, the global "Total Serious and Critical Cases" graph shows a very broad wave, with a quick climb to a peak on April 29 and a slow descent, but within that the numbers go up and down with seemingly no order.  That's what I would expect.  On the Johns Hopkins site, the New Jersey daily numbers show a broad wave with a lot of day-to-day chaos.  New York seems to be settling into a regular diminishing wave pattern.  Texas shows a broadly growing wave with a lot of daily randomness.  That's the pattern I would expect.  Does anyone more familiar with numbers know why such a large scale event would follow a pattern of waves?  Should it be more random?  Am I misreading the graphs and seeing more order than what exists?

It's probably best to use a 7-day rolling average. On some days of the week, fewer tests are released, and not as many counties report the number of deaths.

Exactly this. Here's an example:

Georgia deaths reported on Sundays:
4/5 – 10
4/12 – 5
4/19 – 14
4/26 – 8
5/3 – 3
5/10 – 5
5/17 – 14
Last 7 weeks total: 59

Georgia deaths reported on Mondays:
4/6 – 18
4/13 – 31
4/20 – 46
4/27 – 59
5/4 – 45
5/11 – 36
5/18 – 36
Last 7 weeks total: 271

Does the virus really kill five times as many people on Mondays? Probably not. It's all about reporting.

[wxfree]

Generally you use running averages to smooth things out.  But this virus seems strange to me.  I'm not a statistician, so maybe there's a reason for this I'm not familiar with, but even on the national scale and global scale, large enough that you'd think the irregularities would smooth out, since not every place has the same irregularities, the daily new cases numbers follow a fairly regular wave pattern with peaks and troughs.  I would expect this would be more random, with an average daily number of X over a weeklong period, and each day staying mostly within a certain percentage of that number, but randomly either higher or lower.  In such a pattern, waves would sometimes happen by chance, but you wouldn't expect waves consistently.  It isn't a perfect wave pattern.  Sometimes it goes up, down, then up.  But it isn't nearly as random as I would expect.

On Worldometer, the global "Total Serious and Critical Cases" graph shows a very broad wave, with a quick climb to a peak on April 29 and a slow descent, but within that the numbers go up and down with seemingly no order.  That's what I would expect.  On the Johns Hopkins site, the New Jersey daily numbers show a broad wave with a lot of day-to-day chaos.  New York seems to be settling into a regular diminishing wave pattern.  Texas shows a broadly growing wave with a lot of daily randomness.  That's the pattern I would expect.  Does anyone more familiar with numbers know why such a large scale event would follow a pattern of waves?  Should it be more random?  Am I misreading the graphs and seeing more order than what exists?

It's probably best to use a 7-day rolling average. On some days of the week, fewer tests are released, and not as many counties report the number of deaths.

Exactly this. Here's an example:

Georgia deaths reported on Sundays:
4/5 – 10
4/12 – 5
4/19 – 14
4/26 – 8
5/3 – 3
5/10 – 5
5/17 – 14
Last 7 weeks total: 59

Georgia deaths reported on Mondays:
4/6 – 18
4/13 – 31
4/20 – 46
4/27 – 59
5/4 – 45
5/11 – 36
5/18 – 36
Last 7 weeks total: 271

Does the virus really kill five times as many people on Mondays? Probably not. It's all about reporting.

The global and US numbers do seem to be mostly on a week-long wave pattern.  I hadn't noticed that, but it could help explain the waves as reporting varies through the week.  Global new infections tend to peak on Friday, and US on Friday and for the past two weeks on Thursday.  Since the US numbers are so high, that would have a large influence the shape of the global pattern, which is another thing I didn't think of before.

The US daily death numbers have a less consistent, but still evident wave pattern, and it's just about weekly, too.  I would think that deaths would be reported more consistently since the actual dates are known, unlike when infections happen.  Still, the low points tend to be on Sundays, which I guess is understandable (although I don't like the idea that the death reporting business takes weekends off).  If they do take weekends off, I would expect something like the Georgia pattern, with high counts on Mondays, but the peaks tend to be on Tuesdays (which is even more upsetting if we're so lazy about reporting that it takes two to three days to report weekend deaths).

[J N Winkler]

The US daily death numbers have a less consistent, but still evident wave pattern, and it's just about weekly, too.  I would think that deaths would be reported more consistently since the actual dates are known, unlike when infections happen.

I haven't dug into the exact reporting chain for COVID-19 deaths, but I am reasonably certain that the day a given person appears in the COVID-19 death tally is usually not the same as the day that appears on his or her death certificate.  And some jurisdictions (e.g., Britain) explicitly do not include certain types of COVID-19 deaths in their daily totals.

[CoreySamson]

What state are you in?

I think Texas is in the wait and see state. For example, my home county has had an increase in active cases nearly each day, however, that may be because it has a high concentration of prisons and I don't think the prisons report recovered cases.

The rate change in new cases appears to be slowly increasing in both my county and the state as a whole, but I think that may be attributed to increased testing. Texas has increased its testing from approx 15000 per day to about 20000-25000 this last week.

However, Texas isn't out of the woods just yet.

[LM117]

What state are you in?

I think Texas is in the wait and see state. For example, my home county has had an increase in active cases nearly each day, however, that may be because it has a high concentration of prisons and I don't think the prisons report recovered cases.

It was the same way in Wayne County in eastern North Carolina. The county now has over 800 cases, most of which is due to an outbreak at a state prison in Goldsboro. Recovered cases weren't being reported until the county Board of Commissioners became concerned that the number of cases was making the county look bad from a PR standpoint. Lo and behold, the local newspaper suddenly reported that most of the cases have recovered.

[wxfree]

For some reason the federal government is still moving immigration prisoners around.  It's like they're trying to spread the virus.  Jails, which are ideal places for viruses to spread, are getting infected federal prisoners and the virus spreads like wildfire.  In one case, the workers who were exposed had to quarantine in the jail so it wouldn't get out into the community.  What was reported is that those prisoners, at least the way Texas counts them, add to the state total but not the county totals, since they were imported from other states and don't reflect the conditions in the community.

[Max Rockatansky]

California's current emergency order is presently under some hot water from the Department of Justice for not allowing churches to begin reopening in Phase 2:

https://www.dailynews.com/2020/05/19/california-order-discriminates-against-churches-u-s-says/

[Roadgeekteen]

California's current emergency order is presently under some hot water from the Department of Justice for not allowing churches to begin reopening in Phase 2:

https://www.dailynews.com/2020/05/19/california-order-discriminates-against-churches-u-s-says/

I mean, if a church is following strict guidelines, than they could reopen, but I'm not sure how many churches will. In Massachusetts many churches are remaining closed even after Baker will allow them to reopen.

[Max Rockatansky]

California's current emergency order is presently under some hot water from the Department of Justice for not allowing churches to begin reopening in Phase 2:

https://www.dailynews.com/2020/05/19/california-order-discriminates-against-churches-u-s-says/

I mean, if a church is following strict guidelines, than they could reopen, but I'm not sure how many churches will. In Massachusetts many churches are remaining closed even after Baker will allow them to reopen.

Considering many churches in rural counties (including the one in the news lately in Butte County) already are in defiance of the State Order I would say quite a few. 

[jakeroot]

How come law enforcement officers are mask less?  They deal with people when they pull over or better yet what about a person they have to transport in their back seat?

Too many lawyers who will use the mask as a claim that the officer gave unclear instructions or asked unclear questions.

If HBO's Watchmen is any indication, masks would also conceal identity. This was intentional to protect the police from being tracked down and attacked/murdered.

It's actually quite humerous to see peace officers in various jurisdictions (locally and worldwide) who have been using masks or other face coverings, since it reminds me a ton of Watchmen, apart from them not being yellow:



Image from KIDK Idaho -- All police in Idaho Falls are now wearing masks.

[hbelkins]

In Kentucky, fully two-thirds of the deaths are those in nursing homes, which indicates that age or other medical issues were indicated comorbidities. There was also a major outbreak in one prison in Western Kentucky. Nursing home-related infections and this one particular prison show that a large number of the cases tend to be concentrated in specific areas where repeated exposure in a closely-confined space is a major contributing factor. The nursing home in one county adjoining mine has seen an outsized number of infections and deaths. The two confirmed cases in another adjoining county were from people who worked at that particular nursing home.

My county is currently one of four in Kentucky that does not yet have a confirmed case. All residents and employees of our local nursing home were tested and all tests came back negative.

We had our largest single-day reported number of deaths yesterday -- 20 -- but even our fear-mongering and dictatorial governor acknowledged that the number was probably skewed due to reporting considerations.

[CoreySamson]

California's current emergency order is presently under some hot water from the Department of Justice for not allowing churches to begin reopening in Phase 2:

https://www.dailynews.com/2020/05/19/california-order-discriminates-against-churches-u-s-says/

I mean, if a church is following strict guidelines, than they could reopen, but I'm not sure how many churches will. In Massachusetts many churches are remaining closed even after Baker will allow them to reopen.

Considering many churches in rural counties (including the one in the news lately in Butte County) already are in defiance of the State Order I would say quite a few.

I think it should be the church's decision when they'll open up. There are some in rural areas that probably would be perfectly fine opening up but their state government is still on lockdown.
On the other hand, there probably are churches that would rather not open up right now, even though their state government says they can.

There are just so many factors in the logistics of churches that I don't think a definitive ban (or opening) for them would work. Nevertheless, they should all still follow social distancing rules.

For example, my church's main congregation is only meeting on Sundays at 9 and 11, with no childcare. Only every other pew will be available, and they will still stream services online just as they have for the past few months. Instead of doing a midweek service, they're letting the youth group use the main sanctuary on Wednesday nights.

My point still stands, though. I think it should be the church's decision when they open up.

[ftballfan]

What state are you in?

I think Texas is in the wait and see state. For example, my home county has had an increase in active cases nearly each day, however, that may be because it has a high concentration of prisons and I don't think the prisons report recovered cases.

It was the same way in Wayne County in eastern North Carolina. The county now has over 800 cases, most of which is due to an outbreak at a state prison in Goldsboro. Recovered cases weren't being reported until the county Board of Commissioners became concerned that the number of cases was making the county look bad from a PR standpoint. Lo and behold, the local newspaper suddenly reported that most of the cases have recovered.

In Michigan, prison cases among inmates aren't counted among the county the prison is in; they're recorded separately as "Michigan Department of Corrections". Cases among employees, however, are reported in the county they live in (for example, about a third of Mason County's cases (and at least one case in a few other counties) are among prison workers at the private prison in Lake County)

[TheHighwayMan3561]

I think a large plurality of people accept that things are closed but will eagerly return immediately when things reopen. The "open everything now"  and "keep everything closed/boycott open business"  groups are just vocal minorities.

[Tonytone]

Why are comments getting deleted? This is like the 10th time this has happened.


iPhone

[hotdogPi]

Why are comments getting deleted? This is like the 10th time this has happened.


iPhone

Political discussion is prohibited unless it relates to roads.

[Tonytone]

Why are comments getting deleted? This is like the 10th time this has happened.


iPhone

Political discussion is prohibited unless it relates to roads.

On a corona virus thread? Eh ok.


iPhone

[Roadgeekteen]

Why are comments getting deleted? This is like the 10th time this has happened.


iPhone

Political discussion is prohibited unless it relates to roads.

On a corona virus thread? Eh ok.


iPhone

You can talking about the virus without getting too political.

[Brandon]

Why are comments getting deleted? This is like the 10th time this has happened.


iPhone

Political discussion is prohibited unless it relates to roads.

On a corona virus thread? Eh ok.

You can talking about the virus without getting too political.

And even closures and the like.  You can say "Governor X is mandating that Y shuts down or reopens".  You cannot say "Governor X sucks donkey balls for his closure/reopen strategy".

[kphoger]

Let's try to steer back away from politics... this is not a roads-related post so the tolerance for politics is lower

(emphasis mine)

[kphoger]

The numbers in Kansas are so unbelievably noisy it is hard to get an accurate picture of trends.  Because we had our first confirmed cases late and locked down early, the daily death numbers are small and lumpy (what statisticians call a "small-numbers problem").  With nursing home and meatpacking plant outbreaks in the picture, the underlying mortality rate fluctuates in time since the virus varies in how it samples the age structure (care home residents are mostly older with comorbidities while meatpackers are young). 

Our good friends were in Meade County a couple of weeks ago, visiting friends.  He said it was really weird what a difference there was from just one county away.

Ford County, which is only #17 in population, has had the highest number of cases in the state.  (Finney County, #11 by population, is now in second place.)  Yet, just one county away, there's almost nothing.  Current numbers are 1403 compared to only 22.

Both Ford and Finney counties are home to huge meat-packing plants.

[Max Rockatansky]

Finally got my Virus check yesterday.  I ended up spending about $1,000 dollars on a new mountain bike and parts.  I'll be field testing it out on Blackrock Road in Sierra National Forest come Sunday it seems.  I haven't had to use my bike rack in forever so that will be interesting getting that thing dusted off again. 

[kalvado]

Finally got my Virus check yesterday.  I ended up spending about $1,000 dollars on a new mountain bike and parts.  I'll be field testing it out on Blackrock Road in Sierra National Forest come Sunday it seems.  I haven't had to use my bike rack in forever so that will be interesting getting that thing dusted off again.

Virus check as a nasal swab or blood test?

[Brandon]

Finally got my Virus check yesterday.  I ended up spending about $1,000 dollars on a new mountain bike and parts.  I’ll be field testing it out on Blackrock Road in Sierra National Forest come Sunday it seems.  I haven’t had to use my bike rack in forever so that will be interesting getting that thing dusted off again.

Virus check as a nasal swab or blood test?

Stimulus check.