General > General Technical Chat
Covid 19 virus
not1xor1:
--- Quote from: Dundarave on March 28, 2020, 01:34:13 am ---
--- Quote from: drussell on March 28, 2020, 01:14:51 am ---
--- Quote from: engrguy42 on March 27, 2020, 09:28:56 pm ---FYI, the reason I focus on deaths is because there are far too many unknowns, IMO, to draw any meaningful conclusions on the ultimate impact based solely on number of confirmed cases. For example, Spain and Germany have vastly different numbers of total deaths to date (4089 vs 253, respectively), but have 56k vs. 42k confirmed cases, respectively.
And in the last 2 weeks the US has had a fairly flat average death rate, which is certainly not directly proportional to the increasing # of confirmed cases.
--- End quote ---
:palm:
--- End quote ---
"Fairly flat" isn't how I'd define the US Covid-19 death rate, given the graphs compiled by these folks:
https://www.worldometers.info/coronavirus/country/us/
(Attachment Link)
--- End quote ---
may be it is just a flat EEG :)
Cerebus:
--- Quote from: Dundarave on March 28, 2020, 01:34:13 am ---
--- Quote from: drussell on March 28, 2020, 01:14:51 am ---
--- Quote from: engrguy42 on March 27, 2020, 09:28:56 pm ---FYI, the reason I focus on deaths is because there are far too many unknowns, IMO, to draw any meaningful conclusions on the ultimate impact based solely on number of confirmed cases. For example, Spain and Germany have vastly different numbers of total deaths to date (4089 vs 253, respectively), but have 56k vs. 42k confirmed cases, respectively.
And in the last 2 weeks the US has had a fairly flat average death rate, which is certainly not directly proportional to the increasing # of confirmed cases.
--- End quote ---
:palm:
--- End quote ---
"Fairly flat" isn't how I'd define the US Covid-19 death rate, given the graphs compiled by these folks:
https://www.worldometers.info/coronavirus/country/us/
(Attachment Link)
--- End quote ---
Are we failing basic maths now? That graph does seem to show a fairly flatish death rate. That's a cumulative deaths graph, death rate is d deaths / dt
not1xor1:
--- Quote from: hendorog on March 28, 2020, 06:10:08 am ---Deaths are easier to count, but of course are delayed too much to be useful, impacted by the demographic and existing health conditions,, and impacted by the capacity of the medical system to treat people (ICU beds etc)
Comparing countries is not easy. Germany is clearly a big outlier. Perhaps they are awesome at testing and contact tracing. And so they have detected many more people with minor infections than other countries.
--- End quote ---
here is an interesting article on the subject (unfortunately for some reason the graphics do not appear in the google translated page)
https://translate.google.it/translate?hl=it&tab=wT&sl=it&tl=en&u=https%3A%2F%2Fwww.corriere.it%2Fsalute%2Fmalattie_infettive%2F20_marzo_27%2Fstudio-ispi-ecco-qual-vera-letalita-covid-19-italia-b95d19cc-7029-11ea-82c1-be2d421e9f6b.shtml
Cerebus:
--- Quote from: imo on March 28, 2020, 08:08:07 am ---
--- Quote from: Someone on March 27, 2020, 11:54:23 pm ---
--- Quote from: imo on March 27, 2020, 06:08:46 pm ---While looking for some info on asthma vs. covid I've come across this picture - NO2 concentration in Europe. Would be interesting to compare covid data with the map..
--- End quote ---
https://www.independent.co.uk/environment/satellite-images-emissions-climate-crisis-coronavirus-europe-map-a9426436.html
Infection rates are unlikely to be the correlation, its more about control measures/restrictions being applied in each area.
--- End quote ---
The number of people in ICUs does show a correlation, it seems..
--- End quote ---
Well it does at any time. Try looking at a control period or two, you'll find a correlation between measures of atmospheric pollution and ICU admissions there too (rising NOx => rising ICU occupancy). What you'll then see if you look further is a fall in NOx emissions recently (because nobody is using the roads) and a rise in ICU admissions (coronavirus infections). If you're looking at correlation from the absolute magnitude of difference from the mean (e.g. using variance like the Pearson correlation coefficient, R, which Excel uses) it will treat both positive and negative effects as correlation. If you're using a method that simply looks for any correlation (like Pearson) you'll find it, you need to use something sensitive to the sign of the difference from the mean before you'll see something half meaningful. Let us not forget the mantra of all scientists when dealing with statistics "Correlation does not imply causation".
hamster_nz:
--- Quote from: Cerebus on March 28, 2020, 09:47:18 am ---Are we failing basic maths now? That graph does seem to show a fairly flatish death rate. That's a cumulative deaths graph, death rate is d deaths / dt
--- End quote ---
I guess it looks sort of flat, if you hold your head on the side...
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version