COVID-19

Covid-19 discussion, bring your own statistics
Post Reply
User avatar
jaap
Fuzzable
Posts: 245
Joined: Mon Nov 11, 2019 2:05 pm
Location: Netherlands
Contact:

Re: COVID-19

Post by jaap » Mon Nov 16, 2020 10:29 am

KAJ wrote:
Mon Nov 16, 2020 9:22 am
The blunt fact is that (at 5:00 p.m. 15 Nov) we don't know how many cases (or deaths, admissions, inpatients) there were on 15th, 14th, 13th November. Our information for a few days earlier is suspect. Plotting earlier cases (or deaths) on those dates doesn't really help at all, it is very likely to mislead others and perhaps even ourselves. The only honest way of plotting data for 15th, 14th, 13th is to wait until that data is available.
The daily death stats in the Netherlands are presented like this:
Overledenen-per-dag-vanaf-31-augustus-2020.png
Overledenen-per-dag-vanaf-31-augustus-2020.png (21.38 KiB) Viewed 532 times

Unfortunately they only update them once a week, and the yellow parts are the differences from the previous week's stats. It makes it very clear none of the numbers are final, but more so for the last few days. I do hope tomorrow's update shows some improvement.

Squeak
Snowbonk
Posts: 458
Joined: Tue Dec 03, 2019 6:27 am

Re: COVID-19

Post by Squeak » Mon Nov 16, 2020 11:06 am

Chris Preston wrote:
Sun Nov 15, 2020 9:59 pm
After months of no cases of community transmission in South Australia, there are 17 cases confirmed in the past 2 days, 15 of them in one extended family. Original source was a returning overseas traveler who infected a worker at one of the quarantine hotels.

This demonstrates just how quickly numbers can grow from a single infected person. Even when an effective track and trace system is in place.

No doubt some complacency has set in in South Australia due to the long period of no community transmission and the relative freedom that has brought. If any new cases arise today, we can expect a tightening of restrictions.
That poor family. Not only must they be worked about each other, I can't imagine what it must be like to have borders close and a state go into lockdown just because your family got sick. :(

I'm heartbroken that my first family visit in a year will probably have to be cancelled and I'm trying to be a grown-up about this being a very small price to pay for most families getting to keep grandma and grandpa alive next year, but, well, it feels harder than I expected.

For reference for non-Australians, the zero tolerance approach here means that state borders are snapping shut on the strength of 17 cases because Australians look at Europe and the Americas with horror. We get to live almost normal lives, so long as we don't want to travel interstate. :/

User avatar
jimbob
After Pie
Posts: 1838
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: COVID-19

Post by jimbob » Mon Nov 16, 2020 11:06 am

KAJ wrote:
Mon Nov 16, 2020 9:22 am
jimbob wrote:
Sun Nov 15, 2020 11:34 pm
Yes, and you can do a 7day average on the reported date and smooth out most of the noise - yes the maximum lag might be longer than a week, but the biggest variation is in day of the week.

Unsurprisingly, when you do that, you see that when rising, the reported date lags the actual date equivalent rolling average
But after distorting the data by associating with the wrong (for almost all purposes) date, that introduces a further, artificial, distortion. It may look better, but it's still wrong!

The blunt fact is that (at 5:00 p.m. 15 Nov) we don't know how many cases (or deaths, admissions, inpatients) there were on 15th, 14th, 13th November. Our information for a few days earlier is suspect. Plotting earlier cases (or deaths) on those dates doesn't really help at all, it is very likely to mislead others and perhaps even ourselves. The only honest way of plotting data for 15th, 14th, 13th is to wait until that data is available.
That's why the ONS data is 11-18 days behind the actual day. That's as close to definitive as we have.

The rest is to give us an earlier idea of trends. In March, waiting for the ONS date of death would have reported the first 5 deaths in week 11 (w/e 13-Mar) but we'd have waited to 24th for that data. By which time 253 deaths had been reported (looking at the ONS weekly data for wk44 and the covid daily registrations tab).
Have you considered stupidity as an explanation

KAJ
Stargoon
Posts: 93
Joined: Thu Nov 14, 2019 5:05 pm
Location: UK

Re: COVID-19

Post by KAJ » Mon Nov 16, 2020 6:56 pm

jaap wrote:
Mon Nov 16, 2020 10:29 am
The daily death stats in the Netherlands are presented like this:

Overledenen-per-dag-vanaf-31-augustus-2020.png


Unfortunately they only update them once a week, and the yellow parts are the differences from the previous week's stats. It makes it very clear none of the numbers are final, but more so for the last few days. I do hope tomorrow's update shows some improvement.
A nice presentation which emphasises the delays in reporting. Bar plots aren't really appropriate for log scales where there isn't a meaningful zero and, in this example, the height of the yellow parts would depend on whether they were above or below the purple - counterintuitive. That said, I've been using bar plots on log scales in this thread :oops:. I''m not very interested in the delay. If I were I'd have to change my current practice of daily completely replacing my data set with data from gov.uk, I'd need to keep previous days data set as well.

With respect to "Unfortunately they only update them once a week", my conversation with jimbob is beginning to make me think I'm over-interested in the currency of data - I really should practice the patience I preach and wait for good data to become available. Alas, I can't resist casting the runes and gazing into entrails for signs of what is to come.

KAJ
Stargoon
Posts: 93
Joined: Thu Nov 14, 2019 5:05 pm
Location: UK

Re: COVID-19

Post by KAJ » Mon Nov 16, 2020 7:33 pm

jimbob wrote:
Mon Nov 16, 2020 11:06 am
KAJ wrote:
Mon Nov 16, 2020 9:22 am
The blunt fact is that (at 5:00 p.m. 15 Nov) we don't know how many cases (or deaths, admissions, inpatients) there were on 15th, 14th, 13th November. Our information for a few days earlier is suspect. Plotting earlier cases (or deaths) on those dates doesn't really help at all, it is very likely to mislead others and perhaps even ourselves. The only honest way of plotting data for 15th, 14th, 13th is to wait until that data is available.
That's why the ONS data is 11-18 days behind the actual day. That's as close to definitive as we have.

The rest is to give us an earlier idea of trends. In March, waiting for the ONS date of death would have reported the first 5 deaths in week 11 (w/e 13-Mar) but we'd have waited to 24th for that data. By which time 253 deaths had been reported (looking at the ONS weekly data for wk44 and the covid daily registrations tab).
Yes, I'm beginning to think that my eagerness to see recent changes has clouded my eyes to the nature of the data. As I understand it (often wrongly):
  • Case data on gov.uk isn't intended for estimation of (changes in) prevalence, it's intended to inform treatment of (potential) cases. As such it isn't representative of the population, it's biased towards those considered more likely to be positive. The ONS case data is intended to be representative, As you say, it is even more delayed than the gov.uk data - but perhaps I should be more patient!
  • Hospital admissions, inpatients, and mechanically ventilated beds are also not intended for estimation of (changes in) prevalance, they're intended to help the NHS manage its capacity. For that purpose currency isn't necessary and data is only published once a week or so.
  • Deaths. I'm really not sure why death data is collected/published except that it would be politically impossible not to. There can be substantial delays between death and publication, preceded by more delays between infection and death, so really not suitable for estimating current (changes in) prevalence.
Overall I'm using the data for a purpose to which it's not suited and quibbling about the representation is really pointing out motes while ignoring beams. I suppose I should follow my own advice and wait until appropriate (ONS) data is available - but I know I won't be able to resist playing with the data :geek: .

User avatar
Bird on a Fire
Light of Blast
Posts: 4366
Joined: Fri Oct 11, 2019 5:05 pm
Location: nadir of brie

Re: COVID-19

Post by Bird on a Fire » Mon Nov 16, 2020 7:47 pm

How does the Zoe symptom study compare with the ONS data in the end - similar patterns?
THINK BIG AND UPEND THE SYSTEM

#ShowYourStripes

User avatar
jimbob
After Pie
Posts: 1838
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: COVID-19

Post by jimbob » Mon Nov 16, 2020 8:00 pm

Bird on a Fire wrote:
Mon Nov 16, 2020 7:47 pm
How does the Zoe symptom study compare with the ONS data in the end - similar patterns?
It suggests a peak on 3rd November,

https://covid.joinzoe.com/data
Have you considered stupidity as an explanation

Chris Preston
Stargoon
Posts: 118
Joined: Tue Nov 12, 2019 8:05 am

Re: COVID-19

Post by Chris Preston » Mon Nov 16, 2020 11:35 pm

Squeak wrote:
Mon Nov 16, 2020 11:06 am
Chris Preston wrote:
Sun Nov 15, 2020 9:59 pm
After months of no cases of community transmission in South Australia, there are 17 cases confirmed in the past 2 days, 15 of them in one extended family. Original source was a returning overseas traveler who infected a worker at one of the quarantine hotels.

This demonstrates just how quickly numbers can grow from a single infected person. Even when an effective track and trace system is in place.

No doubt some complacency has set in in South Australia due to the long period of no community transmission and the relative freedom that has brought. If any new cases arise today, we can expect a tightening of restrictions.
That poor family. Not only must they be worked about each other, I can't imagine what it must be like to have borders close and a state go into lockdown just because your family got sick. :(

I'm heartbroken that my first family visit in a year will probably have to be cancelled and I'm trying to be a grown-up about this being a very small price to pay for most families getting to keep grandma and grandpa alive next year, but, well, it feels harder than I expected.

For reference for non-Australians, the zero tolerance approach here means that state borders are snapping shut on the strength of 17 cases because Australians look at Europe and the Americas with horror. We get to live almost normal lives, so long as we don't want to travel interstate. :/
It will be hard on them, but at the moment most people seem to be supportive of getting past this and back to less restrictions. Nobody wants to be where Victoria was. I haven't been on social media yet, so there could be a lot of nonsense there. I am disappointed that one of the media outlets released information about the nationality of the family, as that could result in some racist attacks.

Some good friends of ours who were supposed to be visiting from interstate have of course cancelled. It is the 4th time that we or they have been forced to cancel getting together. Mrs P has not seen her family in almost 2 years and we don't know when that is going to change.

I think there has been plenty to learn from how the pandemic has been managed in Australia. The idea of letting the disease safely bubble along in society at low numbers just doesn't seem possible. It is far to infectious for that and the economic damage of managing large spikes in disease is too high. The whole idea that everyone healthy can catch it and at the same time you can keep it away from vulnerable people (like nursing homes) is just bonkers. While we have gone immediately to new restrictions: gyms and other places closed, restaurants and bars with limited numbers, reduced sizes of gatherings (our annual Thanksgiving dinner has just been cut from 30 to 10) and border closures, we will likely only have to endure this for 2 weeks and then everyone can get back to what they were doing. Allowing numbers to rise and then employing a full-blown lockdown would cost a lot more and last a lot longer, as was seen in Victoria.

Trust you are keeping well and there is a chance South Australia could be back open for business by Christmas.
Here grows much rhubarb.

Squeak
Snowbonk
Posts: 458
Joined: Tue Dec 03, 2019 6:27 am

Re: COVID-19

Post by Squeak » Tue Nov 17, 2020 2:26 am

Chris Preston wrote:
Mon Nov 16, 2020 11:35 pm

It will be hard on them, but at the moment most people seem to be supportive of getting past this and back to less restrictions. Nobody wants to be where Victoria was. I haven't been on social media yet, so there could be a lot of nonsense there. I am disappointed that one of the media outlets released information about the nationality of the family, as that could result in some racist attacks.

Some good friends of ours who were supposed to be visiting from interstate have of course cancelled. It is the 4th time that we or they have been forced to cancel getting together. Mrs P has not seen her family in almost 2 years and we don't know when that is going to change.

I think there has been plenty to learn from how the pandemic has been managed in Australia. The idea of letting the disease safely bubble along in society at low numbers just doesn't seem possible. It is far to infectious for that and the economic damage of managing large spikes in disease is too high. The whole idea that everyone healthy can catch it and at the same time you can keep it away from vulnerable people (like nursing homes) is just bonkers. While we have gone immediately to new restrictions: gyms and other places closed, restaurants and bars with limited numbers, reduced sizes of gatherings (our annual Thanksgiving dinner has just been cut from 30 to 10) and border closures, we will likely only have to endure this for 2 weeks and then everyone can get back to what they were doing. Allowing numbers to rise and then employing a full-blown lockdown would cost a lot more and last a lot longer, as was seen in Victoria.

Trust you are keeping well and there is a chance South Australia could be back open for business by Christmas.
Oh dear. I had assumed that with such a large family and the suburb they live in that there was a decent chance that they'd be non-white and therefore vulnerable to that sort of sh.t. I can't imagine what it must be like to have your whole family sick and to feel as though you can't trust your community to support you.

As disappointed as I am to probably miss my family this year, I cannot begin to express my gratitude that Australia has not followed the callous path of Europe and the US. A short, sharp restriction on activities will almost certainly stop the Adelaide outbreak, and if it doesn't, then a longer lockdown like Melbourne's will do so. And then we can get back to normal. All those boring lesson from 1919 hold - do the lockdown hard at the beginning and it will cost so much less in the longterm, socially and financially.

I guess, with a vaccine on the horizon, it's too late to hope that other countries could learn from our and NZ's experiences though - NZ has had a truly zero tolerance approach and Victoria has demonstrated that you can cut back from fairly widespread community transmission to zero new cases. And they've both done it in English-speaking western countries populated by "people like us", in contrast to SE Asia where some mystical communitarian ideas and previous pandemic experience can be used to explain away compliance with public health directives. This sh.t really is doable, even if I have to miss out on in-person storytime and teaching my niece the finer points of slip'n'slide technique...

User avatar
Martin_B
Snowbonk
Posts: 597
Joined: Mon Nov 11, 2019 1:20 pm
Location: Perth, WA

Re: COVID-19

Post by Martin_B » Tue Nov 17, 2020 4:12 am

Squeak wrote:
Tue Nov 17, 2020 2:26 am
Chris Preston wrote:
Mon Nov 16, 2020 11:35 pm

It will be hard on them, but at the moment most people seem to be supportive of getting past this and back to less restrictions. Nobody wants to be where Victoria was. I haven't been on social media yet, so there could be a lot of nonsense there. I am disappointed that one of the media outlets released information about the nationality of the family, as that could result in some racist attacks.

Some good friends of ours who were supposed to be visiting from interstate have of course cancelled. It is the 4th time that we or they have been forced to cancel getting together. Mrs P has not seen her family in almost 2 years and we don't know when that is going to change.

I think there has been plenty to learn from how the pandemic has been managed in Australia. The idea of letting the disease safely bubble along in society at low numbers just doesn't seem possible. It is far to infectious for that and the economic damage of managing large spikes in disease is too high. The whole idea that everyone healthy can catch it and at the same time you can keep it away from vulnerable people (like nursing homes) is just bonkers. While we have gone immediately to new restrictions: gyms and other places closed, restaurants and bars with limited numbers, reduced sizes of gatherings (our annual Thanksgiving dinner has just been cut from 30 to 10) and border closures, we will likely only have to endure this for 2 weeks and then everyone can get back to what they were doing. Allowing numbers to rise and then employing a full-blown lockdown would cost a lot more and last a lot longer, as was seen in Victoria.

Trust you are keeping well and there is a chance South Australia could be back open for business by Christmas.
Oh dear. I had assumed that with such a large family and the suburb they live in that there was a decent chance that they'd be non-white and therefore vulnerable to that sort of sh.t. I can't imagine what it must be like to have your whole family sick and to feel as though you can't trust your community to support you.

As disappointed as I am to probably miss my family this year, I cannot begin to express my gratitude that Australia has not followed the callous path of Europe and the US. A short, sharp restriction on activities will almost certainly stop the Adelaide outbreak, and if it doesn't, then a longer lockdown like Melbourne's will do so. And then we can get back to normal. All those boring lesson from 1919 hold - do the lockdown hard at the beginning and it will cost so much less in the longterm, socially and financially.

I guess, with a vaccine on the horizon, it's too late to hope that other countries could learn from our and NZ's experiences though - NZ has had a truly zero tolerance approach and Victoria has demonstrated that you can cut back from fairly widespread community transmission to zero new cases. And they've both done it in English-speaking western countries populated by "people like us", in contrast to SE Asia where some mystical communitarian ideas and previous pandemic experience can be used to explain away compliance with public health directives. This sh.t really is doable, even if I have to miss out on in-person storytime and teaching my niece the finer points of slip'n'slide technique...
The irony of this happening just as Western Australia was unlocking our borders with the other states (well, not NSW or Vic yet). NSW seems to be having difficulty stopping all local transmission, with little clusters appearing sporadically, stopped by track-and-trace and self-isolation, but not eradicating it (although there doesn't seem to be any clusters this week!)

I'm tempted by a Christmas holiday in Tassie, but at the moment there's only one flight direct from Perth to Hobart, it's horribly expensive, and I'm not flying via Sydney (which isn't much cheaper!)
"My interest is in the future, because I'm going to spend the rest of my life there"

User avatar
jimbob
After Pie
Posts: 1838
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: COVID-19

Post by jimbob » Tue Nov 17, 2020 1:26 pm

The latest ONS data for England and Wales for Wk 45 (w/e 6th Nov) compared to the last 20 years.

Image

In case you still come across the claim that there are no excess deaths.

COVID-19 is a factor in about 1 in 6 cases and still an almost perfect exponential growth up to Wk 45

Image
Have you considered stupidity as an explanation

bagpuss
Snowbonk
Posts: 505
Joined: Tue Nov 12, 2019 12:10 pm

Re: COVID-19

Post by bagpuss » Tue Nov 17, 2020 1:29 pm

Apologies if this has been posted before and I've missed it but I rather enjoyed this demonstration of the effect of apparently small changes in the R number - Crocheting the R Number

User avatar
sTeamTraen
Dorkwood
Posts: 1211
Joined: Mon Nov 11, 2019 4:24 pm
Location: Palma de Mallorca, Spain

Re: COVID-19

Post by sTeamTraen » Tue Nov 17, 2020 2:26 pm

England has now been in its "second lockdown" for almost two weeks, but unless something dramatic happens in today's figures, UK new cases are not declining --- if anything they're going up. They have been over 20,000 for 28 days now, with only two exceptions. The 7-day moving average is up from 22,700 a week abo to 25,300 as of yesterday evening.

Compare this with Israel, Ireland, Belgium, or the Netherlands, where cases started to fall almost immediately when their second lockdowns started. France is also going the same way, after only a few more days in lockdown than the UK; the 7-day moving average of new cases was 26,000 in the last week compared to 48,000 in the week before that.

This could mean that the new lockdown is having little effect (e.g., if pubs weren't the main cause of transmission, but schools are; on the other hand, France has also shut bars and kept schools open). But what I find strange about the UK numbers is how long they have been bumping along at 20-25,000 (whereas France had a peak of over 80,000).

My worry is that the UK may have hit some intrinsic limit to its test capacity (which is ostensibly large enough), so that the 20-25,000 numbers of a couple of weeks ago were in fact far higher than that. We know there will always be some under-estimation, but the long plateau for the UK numbers seems to be quite unique among comparable countries.
Something something hammer something something nail

User avatar
lpm
After Pie
Posts: 2139
Joined: Mon Nov 11, 2019 1:05 pm
Location: COUNT THE VOTE. STOP THE COUNT.

Re: COVID-19

Post by lpm » Tue Nov 17, 2020 2:55 pm

A 10 day lag isn't too surprising but 2 weeks is starting to push it.

Could be distorted by different regions - the definite fall in former Tier 3 being offset by a general rise elsewhere.

For context

- the SAGE paper that basically called for lockdown and was ignored was on 21 Sept
- the Wales fire break was announced on 19 Oct
- UK lockdown announced 31 Oct

UK official cases on these dates were approx:

21 Sept - 3,700
19 Oct - 17,000
31 Oct - 22,700

The Wales fire break came too late, but I'd hope to see cases below the 17,000 mark next week. Maybe 13,000 a day by 2 Dec? Problem is there's no chance they'll be down to the 3,700 level. It'll be a disaster if we start again with slow exponential growth from 13,000 on 2 Dec.
I'll miss him after he's fled to Riyadh


User avatar
shpalman
After Pie
Posts: 2290
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: COVID-19

Post by shpalman » Tue Nov 17, 2020 7:05 pm

shpalman wrote:
Tue Nov 17, 2020 4:37 pm
Hull is at 770 cases per 100,000
It took me a while to work out that this is based on the number of new cases per week, which is currently 2033, in a population of about 262,000.

It's not huge compared to what's going on in Lombardy.

Still, I didn't see many people wearing masks properly on the news just now.
molto tricky

KAJ
Stargoon
Posts: 93
Joined: Thu Nov 14, 2019 5:05 pm
Location: UK

Re: COVID-19

Post by KAJ » Tue Nov 17, 2020 8:22 pm

sTeamTraen wrote:
Tue Nov 17, 2020 2:26 pm
England has now been in its "second lockdown" for almost two weeks, but unless something dramatic happens in today's figures, UK new cases are not declining --- if anything they're going up. They have been over 20,000 for 28 days now, with only two exceptions. The 7-day moving average is up from 22,700 a week abo to 25,300 as of yesterday evening.
The delayed gov.uk figures have just come out. For reasons discussed earlier I've used cases by specimen date and omitted the most recent 4 days, then the preceding 28 days gives me:
Rplot.png
Rplot.png (17.04 KiB) Viewed 170 times

Code: Select all

Residuals:
      Min        1Q    Median        3Q       Max 
-0.076920 -0.034599 -0.009696  0.017554  0.131704 

Coefficients:
                          Estimate Std. Error t value Pr(>|t|)    
(Intercept)             -1.181e+02  2.610e+01  -4.525 0.000207 ***
date                     6.901e-03  1.406e-03   4.909 8.47e-05 ***
weekdays(date)Monday     2.536e-01  4.154e-02   6.103 5.78e-06 ***
weekdays(date)Saturday  -2.832e-01  4.202e-02  -6.740 1.48e-06 ***
weekdays(date)Sunday    -3.012e-01  4.176e-02  -7.214 5.55e-07 ***
weekdays(date)Thursday   4.233e-02  4.119e-02   1.028 0.316331    
weekdays(date)Tuesday    1.349e-01  4.138e-02   3.260 0.003920 ** 
weekdays(date)Wednesday  1.063e-01  4.126e-02   2.576 0.018022 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.05821 on 20 degrees of freedom
Multiple R-squared:  0.9477,	Adjusted R-squared:  0.9294 
F-statistic: 51.79 on 7 and 20 DF,  p-value: 1.917e-11
Increasing, but very slowly - I estimate a doubling period of about 100 days.

KAJ
Stargoon
Posts: 93
Joined: Thu Nov 14, 2019 5:05 pm
Location: UK

Re: COVID-19

Post by KAJ » Tue Nov 17, 2020 10:35 pm

In case anyone is interested, here is deaths by date of death for the same dates as my last post.
000007.png
000007.png (9.79 KiB) Viewed 138 times

Code: Select all

Residuals:
      Min        1Q    Median        3Q       Max 
-0.155991 -0.053376  0.004689  0.045044  0.222391 

Coefficients:
               Estimate Std. Error t value Pr(>|t|)    
(Intercept)     5.65685    0.01525 370.872  < 2e-16 ***
poly(date, 2)1  1.25922    0.08071  15.602 2.15e-14 ***
poly(date, 2)2 -0.50569    0.08071  -6.265 1.48e-06 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.08071 on 25 degrees of freedom
Multiple R-squared:  0.9187,	Adjusted R-squared:  0.9122 
F-statistic: 141.3 on 2 and 25 DF,  p-value: 2.362e-14
Here I've added a quadratic term because it substantially helped the fit and omitted the weekdays term because it didn't.

I am not convinced that the apparent decline in the last four points is real - I'm ready to believe that the delay between deaths and publication is greater than that between test specimens and publication.

User avatar
jimbob
After Pie
Posts: 1838
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: COVID-19

Post by jimbob » Tue Nov 17, 2020 10:42 pm

KAJ wrote:
Tue Nov 17, 2020 8:22 pm
sTeamTraen wrote:
Tue Nov 17, 2020 2:26 pm
England has now been in its "second lockdown" for almost two weeks, but unless something dramatic happens in today's figures, UK new cases are not declining --- if anything they're going up. They have been over 20,000 for 28 days now, with only two exceptions. The 7-day moving average is up from 22,700 a week abo to 25,300 as of yesterday evening.
The delayed gov.uk figures have just come out. For reasons discussed earlier I've used cases by specimen date and omitted the most recent 4 days, then the preceding 28 days gives me:
Rplot.png

Code: Select all

Residuals:
      Min        1Q    Median        3Q       Max 
-0.076920 -0.034599 -0.009696  0.017554  0.131704 

Coefficients:
                          Estimate Std. Error t value Pr(>|t|)    
(Intercept)             -1.181e+02  2.610e+01  -4.525 0.000207 ***
date                     6.901e-03  1.406e-03   4.909 8.47e-05 ***
weekdays(date)Monday     2.536e-01  4.154e-02   6.103 5.78e-06 ***
weekdays(date)Saturday  -2.832e-01  4.202e-02  -6.740 1.48e-06 ***
weekdays(date)Sunday    -3.012e-01  4.176e-02  -7.214 5.55e-07 ***
weekdays(date)Thursday   4.233e-02  4.119e-02   1.028 0.316331    
weekdays(date)Tuesday    1.349e-01  4.138e-02   3.260 0.003920 ** 
weekdays(date)Wednesday  1.063e-01  4.126e-02   2.576 0.018022 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.05821 on 20 degrees of freedom
Multiple R-squared:  0.9477,	Adjusted R-squared:  0.9294 
F-statistic: 51.79 on 7 and 20 DF,  p-value: 1.917e-11
Increasing, but very slowly - I estimate a doubling period of about 100 days.
It does look as though the exponential rise in deaths is slowing down. Still a bit early to be certain though
Screenshot 2020-11-17 223602.png
Screenshot 2020-11-17 223602.png (37.83 KiB) Viewed 134 times
Have you considered stupidity as an explanation

User avatar
sTeamTraen
Dorkwood
Posts: 1211
Joined: Mon Nov 11, 2019 4:24 pm
Location: Palma de Mallorca, Spain

Re: COVID-19

Post by sTeamTraen » Wed Nov 18, 2020 12:05 am

On 4 November I wrote to the Royal College of Pathologists to point out that a frootloop with the Twitter handle @ClareCraigPath was using the screen name "Dr Clare Craig FRCPath" and asking if they could maybe suggest to her that she shouldn't use her professional qualification/membership to promote frootloopery (most notably, "almost all COVID cases are false positives, death and ICU admissions are flu").

Today I notice that her screen name is now just "Dr Clare Craig" and her bio includes the sentence "Views my own not the RCPath's."

I may very well not have been the only person to complain about her (I didn't get a reply to my e-mail), but this feels like a small victory.
Something something hammer something something nail

Chris Preston
Stargoon
Posts: 118
Joined: Tue Nov 12, 2019 8:05 am

Re: COVID-19

Post by Chris Preston » Wed Nov 18, 2020 2:49 am

South Australia has now gone into a full lockdown. Not allowed to leave home, even for exercise.

This will be fun.
Here grows much rhubarb.

Squeak
Snowbonk
Posts: 458
Joined: Tue Dec 03, 2019 6:27 am

Re: COVID-19

Post by Squeak » Wed Nov 18, 2020 3:16 am

Chris Preston wrote:
Wed Nov 18, 2020 2:49 am
South Australia has now gone into a full lockdown. Not allowed to leave home, even for exercise.

This will be fun.
I'm thinking of you, Chris. Apparently the supermarkets have gone nuts in one last burst of panic-spreading before the lockdown starts. This week is going to brutal for Adelaideans.

(For non-Aussies, things are happening fast so I won't link to a specific news story, but this is lockdown to a degree that I don't think we've seen outside China yet. Everything is closed, apart from essential infrastructure and supermarkets. But with a short generation time, a week of lockdown might just be enough.)

I've seen the reporting around the Adelaide strain having a generation time of three days, with it taking 24 hours from exposure to infectiousness, but I don't have good comparison figures for "normal" covid. Does anyone know if I am right in thinking that it's about 5-6 days?

User avatar
Martin_B
Snowbonk
Posts: 597
Joined: Mon Nov 11, 2019 1:20 pm
Location: Perth, WA

Re: COVID-19

Post by Martin_B » Wed Nov 18, 2020 4:05 am

Chris Preston wrote:
Wed Nov 18, 2020 2:49 am
South Australia has now gone into a full lockdown. Not allowed to leave home, even for exercise.

This will be fun.
Good luck Chris. If you feel the need to vent, please let forth.
"My interest is in the future, because I'm going to spend the rest of my life there"

Chris Preston
Stargoon
Posts: 118
Joined: Tue Nov 12, 2019 8:05 am

Re: COVID-19

Post by Chris Preston » Wed Nov 18, 2020 5:57 am

Yeah it has been a bit hectic this afternoon.

Had to contact my team and organise them to get as much work they could take home that was safe for them to do. I now have a large number of pots of plants being looked after in my back yard. Make contingency plans for all the things that could go wrong while everyone was at home. Check that the postgraduate students who are trying to get submitted next month have all the support they need. Held an exam preparation session for the undergraduate students over Zoom and reassured them that there would be technical help for them to complete the no-line exam this week. Cancel our Thanksgiving dinner. First time in 30 years of living in Australia we won't put it on. Fortunately the turkeys hadn't been killed yet, so they get a reprieve for a while.

The supermarkets are a mess, although our local one wasn't too bad except for the massive queue to pay and the lack of toilet paper and bread. The lines at the bottle shops are absolutely huge as everyone tries to stock up on booze. The roads are just a car park everywhere as everyone rushes out to do things they now won't be able to do for a week.

Fortunately, I have a backlog of writing to do. There is that book chapter due at the end of the year that I have barely started on.
Here grows much rhubarb.

Squeak
Snowbonk
Posts: 458
Joined: Tue Dec 03, 2019 6:27 am

Re: COVID-19

Post by Squeak » Wed Nov 18, 2020 6:30 am

Good luck with the mad scramble, Chris. And may I congratulate you on your procrastination skills - it was very clever of you to plan all year for a quiet week at home in November.

In my immediate Adelaide family, one Kelpie, two German shorthaired pointers, and one hearthrug are about to go unexercised for a week. I am very glad that I don't have to share a house with any of them.

I'm increasingly puzzled by the generation time though. I'll be interested to see if the really quick infection rate is related to the environment rather than something genetic. Most of the patients are from a single family, perhaps they were uniquely vulnerable to a quick incubation. It seems more likely than there being a mysterious new speedy strain of the virus.

Post Reply