Geekzone: technology news, blogs, forums
Guest
Welcome Guest.
You haven't logged in yet. If you don't have an account you can register now.


DS248

1702 posts

Uber Geek
+1 received by user: 771

Lifetime subscriber

#268639 30-Mar-2020 02:32
Send private message

Starting a new thread for data, analysis, research links & related discussion as the frequency of posts to the main COVID-19 thread makes it difficult to locate info that was posted even a day of two ago.

 

==

 

COVID-19 case growth trends plot updated to 28 March 2020. 

 

Have split into separate plots to focus on specific issues.

 

First a few notes & qualifications.

 

  • Day 0 in plot = the day when the number of cases was closest to 150. In the most cases the day 0 number of cases was within the range of 150±10, with the main outliers being the US (175), China (121), and South Korea (104)
  • Data plotted are the absolute numbers of confirmed cases reported mainly by either Wikipedia or Worldometers.com
  • The numbers of confirmed cases are subject to the extents and criteria of testing which vary between significantly jurisdictions and within jurisdictions, over time
  • The numbers of confirmed cases will in most cases be less than the numbers of symptomatic infections due to both a lag between onset of symptoms and confirmation of infection and because restrictive criteria for testing mean many symptomatic people are not tested and hence not counted.  In some jurisdictions though confirmed cases include people who were asymptomatic at time of testing.
  • No adjustment has been made for differences in the populations of various countries.
  • The initial rate of exponential growth in cases above ~150 is mainly a function of the nature of infection spread of the virus and various socio-economic and behavioural factors, and is largely independent of total population. Timing and extent of roll-off from exponential growth is dependent on containment measures and other restrictions implemented
  • ...

 

 

==

 

Plot 1:  East Asian vs Western countries

 

‘High Income’ jurisdictions only (per World Bank ranking + Taiwan)

 

Masks & other factors?

 

 

 

 

As per plot, the rates of growth in COVID-19 cases in Taiwan, Singapore, Hong Kong & Japan follow a very similar trend and are much lower than in the US and Europe.   Twenty days after reaching 150 cases, Germany & Italy had 23 times as many cases as Singapore & Japan, and the US had about 80 times as many cases at the same stage.  At least until two days after lockdown, NZ was clearly following the trend in US and Europe.

 

None of the four Asian countries have yet enforced mass business closures though in recent days restrictions do appear to be increasing. Tokyo subway trains & streets are still crowded, offices in the CBD still working.  

 

In addition to rapidly implement widespread testing and stringent self isolation/quarantine requirements on infected individuals (possibly less so in Japan?), one notable difference is wearing of face masks in the East Asian countries.

 

One issue for these countries is that though at a lower rate, near exponential growth in cases is still continuing.  Not sustainable in the longer term.  Suggests that more aggressive measures are needed and in the last few days more stringent quarantine requirements on travellers have been implemented at least in some of the countries, along with requests for more social distancing and avoiding large gatherings.  Will probably be 7 – 10 days before the effects of these measures are evident.

 

South Korea is different as they had an early large cluster of COVID-19 cases associated with a religious sect.  But with most people wearing face masks, widespread testing, stringent self isolation/quarantine requirements, and school closures, they managed to bring a runaway growth in cases largely under control within about two weeks without mass business closures.  The rate of increase in cases in South Korea is now lower than in Taiwan, Singapore, Hong Kong or Japan.

 

https://www.goodhousekeeping.com/health/wellness/a31820433/coronavirus-south-korea-essay-quarantine/

 

https://www.voanews.com/science-health/coronavirus-outbreak/south-koreas-coronavirus-plan-working-can-world-copy-it

 

 


Filter this topic showing only the reply marked as answer View this topic in a long page with up to 500 replies per page Create new topic
 1 | 2 | 3 | 4 | 5 | 6 | ... | 21
DS248

1702 posts

Uber Geek
+1 received by user: 771

Lifetime subscriber

  #2449825 30-Mar-2020 03:53
Send private message

Plot 2:  Seasonality?

 

The question of seasonality has been speculated on quite a bit, especially in heavily impacted northern hemisphere countries.

 

Plot compares rates of growth for western countries in northern hemisphere (winter & early spring), with those for higher income southern hemisphere countries (summer & early autumn).  Trend for Japan has been retained for contrast.

 

  

 

 

 

 

Plot does suggest a possible impact of season (winter vs summer) but the effect if real, appears to be comparatively small - at least compared to the difference between western countries in the northern hemisphere and the East Asian countries.  The growth rate trends for Chile, Argentina & Australia are quite similar, and are lower than those for the northern hemisphere countries.  The number of cases in Australia at 'Day 16' (28 Mar) is about half the number in Germany and France at the same time.

 

UK has been excluded from the plot due to their quite restrictive testing criteria which lowers the number of confirmed cases compared to some other countries (apparent rate similar to Australia).  I could not find info on numbers of tests in Chile but the number for Australia a day or two ago was ~178,000.  On a per capita basis that is similar to South Korea, and about double the per capita rate for NZ to date.  Also have not included South Africa due potential socio-economic & other impacts.  Their rate appears to be intermediate between the northern and other southern countries.

 

Rate of increase in NZ was initially in line with western countries in the northern hemisphere but has reduced in the last two days.  The numbers here now are similar to the other southern countries at the same stage.  As noted in a previous post, if our cases are heavily import linked, roll off from exponential would likely be evident within 4 or 5 days of lockdown as some people arrive from overseas with symptoms and show up in the stats within a couple of days (+ plus the reduction in people arriving from overseas over the last few days).  Otherwise, it will probably take 7 - 10+ days for the lockdown to significantly impact the growth in numbers of cases.  However, it is still too early to know whether the lockdown and tightened border have contributed to the decrease in rate here in the last two days, or whether the reduction in new daily cases is simply random scatter.  

 

 

 

 

 

 

 

 

 

 




Tinkerisk
4798 posts

Uber Geek
+1 received by user: 3660


  #2449827 30-Mar-2020 04:20
Send private message

Problem I see in these graphs is still, that you have "confirmed cases" on display. As experience shows even here in Germany with a very high dense testing, you don't have the truth with the unknown cases out there. So there is plenty room to underestimate what's really going on.

 

Question is: Are numbers for a specific country high due to close testing or are they high 'cause spreading was out of control the time before?





- NET: FTTH & VDSL, OPNsense, 10G backbone, GWN APs
- SRV: 12 RU HA server cluster, 0.1 PB storage on premise
- IoT:   thread, zigbee, tasmota, BidCoS, LoRa, WX suite, IR
- 3D:    two 3D printers, 3D scanner, CNC router, laser cutter


davidcole
6099 posts

Uber Geek
+1 received by user: 1465

Trusted

  #2449833 30-Mar-2020 07:11
Send private message

Add the other countries lock downs.   I started trying to push all the data to influx since I couldn’t put get excel to do it.  Just to see what the lag and curve looks like.





Previously known as psycik

Home Assistant: Gigabyte AMD A8 Brix, Home Assistant with Aeotech ZWave Controller, Raspberry PI, Wemos D1 Mini, Zwave, Shelly Humidity and Temperature sensors
Media:Chromecast v2, ATV4 4k, ATV4, HDHomeRun Dual
Server
Host Plex Server 3x3TB, 4x4TB using MergerFS, Samsung 850 evo 512 GB SSD, Proxmox Server with 1xW10, 2xUbuntu 22.04 LTS, Backblaze Backups, usenetprime.com fastmail.com Sharesies Trakt.TV Sharesight 




Shadowfoot
First time caller
399 posts

Ultimate Geek
+1 received by user: 256

Trusted
Lifetime subscriber

  #2449867 30-Mar-2020 09:03
Send private message

I thought this was an interesting approach to take. https://aatishb.com/covidtrends/

 

Both scales are logs and show the increase in cases compared with current cases, with time being represented by the animation. One limiting factor which does help smooth out the line, is it's using something like a moving average over a week, so changes will take longer before they appear.  





Dingbatt
6804 posts

Uber Geek
+1 received by user: 3694

Lifetime subscriber

  #2449876 30-Mar-2020 09:43
Send private message

What is the criteria used to say a person has recovered? Is it 2 weeks clear of symptoms?

 

If that is the case then it would be handy to know how many of those initial batch are still ill because a graph of infections can only ever plateau vs a graph of those with the illness still active can show a negative slope. With NZers still pouring back into the country, it would also be great to see graphs for community transmission vs imported (+ close contact). Although showing these numbers may be counterproductive if they show us ‘winning’ too soon.





“We’ve arranged a society based on science and technology, in which nobody understands anything about science technology. Carl Sagan 1996


JWR

JWR
821 posts

Ultimate Geek
+1 received by user: 272


  #2449911 30-Mar-2020 10:22

I find this graph interesting Covid Trends

 

It is a log/log type of graph of new infections to total cases.

 

This gives better view of how each country is doing.

 

Large countries don't overshadow smaller countries on the graph.

 

You can clearly see when a country drops off from exponential growth of the infection (e.g. China, South Korea).

 

Note: You have to manually select New Zealand to play (play button at bottom left of graph) our progress.

 

 


 
 
 
 

Shop now for Lenovo laptops and other devices (affiliate link).
DS248

1702 posts

Uber Geek
+1 received by user: 771

Lifetime subscriber

  #2449976 30-Mar-2020 12:00
Send private message

Quick correction.  In my first post above I said the "rate of increase in cases in South Korea is now lower than in Taiwan, Singapore, Hong Kong or Japan".  In terms of absolute numbers, except for Japan, that is incorrect (sorry, was posted in a bit of a rush in the silly hours of the morning & was fooled by the log scale).  On a per capita basis, the increase in daily numbers in Taiwan and Japan is lower than in South Korea.  


DS248

1702 posts

Uber Geek
+1 received by user: 771

Lifetime subscriber

  #2449985 30-Mar-2020 12:20
Send private message

Agree, all valid points. One of my sons had pointed me to the graph of new infections vs total cases and I agree it better highlights the drop off in cases when countries get on top of the situation. 

 

I have been intending to plot new cases separately as I was conscious the semi-log plot of total cases does not highlight the significant drop off in cases in places like South Korea & China.  A question of time! 

 

Theoretically, I am still working full time but have been a bit distracted this last week.  Having worked from home for almost three decades, the last two almost solely for overseas clients, little has changed work-wise for me (other than the distraction).  Time to get back to work (this afternoon, perhaps)!  I suspect right now NZ is going to need all the overseas income it can get.


DS248

1702 posts

Uber Geek
+1 received by user: 771

Lifetime subscriber

  #2450013 30-Mar-2020 12:39
Send private message

davidcole: Add the other countries lock downs. ...

 

Something I had intended doing for Italy when it was clear they started rolling off exponential.  Slight complication as their lockdown was implemented in stages.  It would be a bit messy to show for many countries on the one plot.  Some things do stand out in the curves; eg. the uptick in US cases once they started more extensive testing.  But, yes it would be good to see what the lag is so I will look at how this can be best shown.  Again, time!  Particularly tracking down lockdown implementation details for a dozen or so countries.  Any links, details etc always a help 😀.


davidcole
6099 posts

Uber Geek
+1 received by user: 1465

Trusted

  #2450025 30-Mar-2020 12:58
Send private message

DS248:

 

davidcole: Add the other countries lock downs. ...

 

Something I had intended doing for Italy when it was clear they started rolling off exponential.  Slight complication as their lockdown was implemented in stages.  It would be a bit messy to show for many countries on the one plot.  Some things do stand out in the curves; eg. the uptick in US cases once they started more extensive testing.  But, yes it would be good to see what the lag is so I will look at how this can be best shown.  Again, time!  Particularly tracking down lockdown implementation details for a dozen or so countries.  Any links, details etc always a help 😀.

 

 

 

 

piss, I thought I'd worked out most of it.  And has added a column with the countries that put in a Nationwide lockdown on the corresponding day.

 

 





Previously known as psycik

Home Assistant: Gigabyte AMD A8 Brix, Home Assistant with Aeotech ZWave Controller, Raspberry PI, Wemos D1 Mini, Zwave, Shelly Humidity and Temperature sensors
Media:Chromecast v2, ATV4 4k, ATV4, HDHomeRun Dual
Server
Host Plex Server 3x3TB, 4x4TB using MergerFS, Samsung 850 evo 512 GB SSD, Proxmox Server with 1xW10, 2xUbuntu 22.04 LTS, Backblaze Backups, usenetprime.com fastmail.com Sharesies Trakt.TV Sharesight 


DS248

1702 posts

Uber Geek
+1 received by user: 771

Lifetime subscriber

  #2450058 30-Mar-2020 13:50
Send private message

Tinkerisk:

 

Problem I see in these graphs is still, that you have "confirmed cases" on display. As experience shows even here in Germany with a very high dense testing, you don't have the truth with the unknown cases out there. So there is plenty room to underestimate what's really going on.

 

Question is: Are numbers for a specific country high due to close testing or are they high 'cause spreading was out of control the time before?

 

 

Yes, it is an issue.  At this stage it is very difficult / not possible to assess in most cases.  There is a bit of info coming out for some areas.  Iceland data, and German & South Korea data give some insight.  A few other areas as well.   Per capita testing in Australia (over 200,000 tests = over 8,000 per million population) is now a bit above South Korea (~7,800 pm).  Plus I have read of work being done China investigating this type of issue.  Fully reliable information is still well down the track.

 

A few things have 'surprised' me a bit though:

 

  • There appears to have been a degree of consistency in test criteria in quite a few countries.  NZ not the only country requiring both symptoms and contact with confirmed or probable case before testing (which for NZ means we do not have reliable info about community transmission).
  • The consistency in growth trends for Singapore, Taiwan, Hong Kong and Japan despite differences in approach (esp. Japan)
  • The consistency in growth trends for Australia, Chile and Argentina.  I know little about measures in the latter two countries but can imagine that they differ somewhat from Australia, a somewhat wealthier country.  Probably need to search in Spanish to find more info, which would be rather time consuming.  Any info would of course be of interest!
  • And Germany and France (cases, not fatalities).  Even Italy (again cases, not fatalities)

One thing I have been waiting to see is where the NZ cases track over the next several days.  Three possibilities of specific interest:

 

     

  1. Continue going down due to early impact of the lockdown (would require imported cases to be a very dominant factor)
  2. Rise again. Drop over the previous two days just a blip
  3. Start tracking along the Australia - Chile - Argentina trend (seasonality effect?)

 

Data for today released since I started this post - 75 new cases (vs 63 yesterday) so back up a little.  Very early days but as at today it is on the line for those countries.  Will need a few more days to know though. 

 

Edit: changed wording in second sentence of last paragraph


 
 
 
 

Shop now for Dell laptops and other devices (affiliate link).
DS248

1702 posts

Uber Geek
+1 received by user: 771

Lifetime subscriber

  #2451889 31-Mar-2020 23:52
Send private message

Interesting comparison between Italy and Germany (data for 24 March).  https://towardsdatascience.com/why-are-covid-19-statistics-so-different-for-germany-and-italy-ee5bf376f461

 

A large disparity between the age distribution of cases in Germany and Italy despite the age distributions of their populations being quite similar.  Given the high fatality rates for over 70 years, that would explain a good fraction of the difference in death rates (with overwhelmed medical facilities in Italy accounting for much of the rest.  Maybe the larger portion?)

 

Back then (24 Mar) the CFR in Germany was 1.6%.  In the week since, the number of cases in Germany has more than doubled and their CFR has risen to 2%.  Possibly more older people are now being affected?

 

 

 

 

 

 

 


frankv
5705 posts

Uber Geek
+1 received by user: 3666

Lifetime subscriber

  #2451923 1-Apr-2020 08:21
Send private message

The key difference between Italy and Germany is that Italy's health system has been swamped.

I think that for comparison between countries, it would be useful to get the deaths and infections per ICU bed, or per capita.

DS248

1702 posts

Uber Geek
+1 received by user: 771

Lifetime subscriber

  #2452187 1-Apr-2020 11:39
Send private message

frankv: The key difference between Italy and Germany is that Italy's health system has been swamped. ...

 

Yes, hence my bit in brackets "(with overwhelmed medical facilities in Italy accounting for much of the rest.  Maybe the larger portion?)"

 

I am trying to assess the relative contributions of the overwhelmed medical facilities and the different age distributions but it is proving difficult & rather time consuming to locate all the necessary data.  But it is clear that the impact of the latter factor (different age distributions) will be far from small given the size of the disparity.


ResponseMediaNZ
518 posts

Ultimate Geek
+1 received by user: 196

ID Verified
Trusted

  #2452341 1-Apr-2020 15:24
Send private message

Word of warning here, can we please only share official information and official research, not a bunch of research out of the back of a truck or I've even seen research that's come out as new but is based on case study 3 months ago


 1 | 2 | 3 | 4 | 5 | 6 | ... | 21
Filter this topic showing only the reply marked as answer View this topic in a long page with up to 500 replies per page Create new topic








Geekzone Live »

Try automatic live updates from Geekzone directly in your browser, without refreshing the page, with Geekzone Live now.



Are you subscribed to our RSS feed? You can download the latest headlines and summaries from our stories directly to your computer or smartphone by using a feed reader.