R
Ricky C
Guest
On Sunday, April 26, 2020 at 6:06:42 AM UTC-4, DecadentLinux...@decadence.org wrote:
Not a bad video. One point that is not clear to me is that the distribution is for mapping the blue line (actual dates of infection) to the orange line (reported dates of infection). He applies it in reverse. It seems to give good results for the one test case, but this is not really a reciprocal relationship. For example, if the blue curve were in reality an impulse function the orange curve would be the binomial distribution he is using. But applying his method to that orange curve would not produce the original blue curve of a single peak.
In reality it is just a coincidence that the nature of the spread of the disease produces a blue curve that "sort of" can be regenerated from the orange curve.
The US CDC does have date of infection data available which I was collecting and plotting to see how the shape of the curve changes with time. A factor that he didn't take into account was reporting delays which we seem to have an excess of in the US. This hugely distorts the data and creates significant waves and peaks in the orange curve (reported dates). The most notable is probably due to weekend effects. There is a pretty clear pattern in reduced numbers Sunday through Tuesday.
I collect the daily data from worldometers.info for the US but do not do the smoothing he does. The last few days have seen a sharp (relatively speaking) uptick in the daily new infection numbers so that the linear regression on data in April is now trending upward again. It had reached a down slope of -150 (about) per day. Now it is up 34 per day.
That does seem to validate the new infection rate of approximately 1.0 he came up with. I suppose that's a DUH since he's using the same raw data.
But it just underlines the fact that in the US we are not doing enough to stem the infection rate and will not be rid of this disease for a long time to come.
This truly is one of those times when a national leader who knows how to influence people and behavior rather than making "deals" would be so much more effective. I mean, Jesus!, he was having verbal brawls with Governors through daily press meetings and encouraging the states to fend for themselves all the while working against them by hording PPE and other goods needed to fight this disease.
Imagine how much more encouraged people would have been had nearly any President been in charge other than the orange baby! GHW Bush would have been very inspirational I think. Anyone else from the last 50 years would have at least not encouraged us to compete and fight.
Ok, end of rant. The message should be that we all need to take this seriously so we can get through it and out the other side.
--
Rick C.
+++ Get 1,000 miles of free Supercharging
+++ Tesla referral code - https://ts.la/richard11209
David Brown <david.brown@hesbynett.no> wrote in news:r83k1j$75o$1@dont-
email.me:
Not good, but not nearly as bad.
Located a nice plot over time for many diseases, so folks can get an
idea of just how bad this one is and is going to be yet. And with the
timeline of growth shown, it should be clear to anyone with an IQ over
20 that we need to remain distant and out of closed space gatherings.
https://www.youtube.com/watch?v=LnQcbAKWkPE
Not a bad video. One point that is not clear to me is that the distribution is for mapping the blue line (actual dates of infection) to the orange line (reported dates of infection). He applies it in reverse. It seems to give good results for the one test case, but this is not really a reciprocal relationship. For example, if the blue curve were in reality an impulse function the orange curve would be the binomial distribution he is using. But applying his method to that orange curve would not produce the original blue curve of a single peak.
In reality it is just a coincidence that the nature of the spread of the disease produces a blue curve that "sort of" can be regenerated from the orange curve.
The US CDC does have date of infection data available which I was collecting and plotting to see how the shape of the curve changes with time. A factor that he didn't take into account was reporting delays which we seem to have an excess of in the US. This hugely distorts the data and creates significant waves and peaks in the orange curve (reported dates). The most notable is probably due to weekend effects. There is a pretty clear pattern in reduced numbers Sunday through Tuesday.
I collect the daily data from worldometers.info for the US but do not do the smoothing he does. The last few days have seen a sharp (relatively speaking) uptick in the daily new infection numbers so that the linear regression on data in April is now trending upward again. It had reached a down slope of -150 (about) per day. Now it is up 34 per day.
That does seem to validate the new infection rate of approximately 1.0 he came up with. I suppose that's a DUH since he's using the same raw data.
But it just underlines the fact that in the US we are not doing enough to stem the infection rate and will not be rid of this disease for a long time to come.
This truly is one of those times when a national leader who knows how to influence people and behavior rather than making "deals" would be so much more effective. I mean, Jesus!, he was having verbal brawls with Governors through daily press meetings and encouraging the states to fend for themselves all the while working against them by hording PPE and other goods needed to fight this disease.
Imagine how much more encouraged people would have been had nearly any President been in charge other than the orange baby! GHW Bush would have been very inspirational I think. Anyone else from the last 50 years would have at least not encouraged us to compete and fight.
Ok, end of rant. The message should be that we all need to take this seriously so we can get through it and out the other side.
--
Rick C.
+++ Get 1,000 miles of free Supercharging
+++ Tesla referral code - https://ts.la/richard11209