Corona Virus and Exponential Growth

@joe-shmo

He is the Khan Academy guy. He drags on a bit, but I thought it was a great explanation.

@Eladar

Yeah, I think he went out on his own. I didn't think this one was that dragging in comparison to some of his others, but I have a feeling it strongly depends on the subject matter/listener engagement.

@joe-shmo said
@Eladar

Yeah, I think he went out on his own. I didn't think this one was that dragging in comparison to some of his others, but I have a feeling it strongly depends on the subject matter/listener engagement.

He does drag on, but he gives good explanations if you are patient enough.


If you struggle with math, you will struggle with his explanations. He does not try to stay basic for others to follow. If you can't keep up, watch it over and over again and it might make sense.

@joe-shmo said
The curve is not purely exponential, its logistical. I think you are going to greatly overestimate.

[youtube COVID19 - Exponential Growth Explained]Kas0tIxDvrg[/youtube]

In fact the curve will go into saturation at some point. The start of a logistical curve is exponential. As the incubation period for Covid-19 is about two weeks I expect at least for the next two weeks further exponential behaviour, when the intervention will be succesfull we will go into the first saturation...after some time restrictions will be lifted. If we don't have a "sumer effect" like with the flu the next wave is to be expected.

@Ponderable

How do you explain the US death rate? Looks like we will be sitting at 50 for a 4th day now.

@eladar said
@Ponderable

How do you explain the US death rate? Looks like we will be sitting at 50 for a 4th day now.

variation. If the counter stops at 50 again and tomorrow we can begin to draw conclusions. Three data points a too weak for that.

Could you give your exponential equation for predicted US deaths, defining unit of time and what you are using for t=0?

Edit ok Feb 29 is day 0 and is measured in days.

Total deaths =1.61*e^(.24t)

@ponderable said
In fact the curve will go into saturation at some point. The start of a logistical curve is exponential. As the incubation period for Covid-19 is about two weeks I expect at least for the next two weeks further exponential behaviour, when the intervention will be succesfull we will go into the first saturation...after some time restrictions will be lifted. If we don't have a "sumer effect" like with the flu the next wave is to be expected.

Your projecting 3 times the current worldwide deaths ( in which some stats have been accumulating for several months, or at least a month in most cases ) to occur in the US alone over the next 3 weeks? I don't buy it.

@joe-shmo

He is projecting over 38 thousand US deaths in 3 weeks.

@joe-shmo
Very informative. Can we see from the C19 data from today if we have reached the inflection point? If not, can we half ass predict when the IP will be reached?

@sonhouse

https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf


Here you will find expert opinion. The death rate curve reaches maximum in June and zeroes in August symmetrically. April to June you see the growth, so somewhere in May the change in death rate should max out. Aka this is where you should find your point of inflection

@sonhouse said
@joe-shmo
Very informative. Can we see from the C19 data from today if we have reached the inflection point? If not, can we half ass predict when the IP will be reached?

"Can we see from the C19 data from today if we have reached the inflection point?"

For that question, the easiest way to see what is happening is to go to: https://www.worldometers.info/coronavirus/country/us/

Scroll down to :Total Coronavirus Cases in the US

On the chart, click the word "logarithmic".
If the slope of that "line" isn't changing ( it looks like y=m*x+b , m>0 ) then we aren't at the inflection point.

Now do the same thing with Italy: https://www.worldometers.info/coronavirus/country/italy/

You can start to see that the "steepness" of that trend is decreasing. It appears to me as though Italy has hit the inflection point.

Finally, do the same for China: https://www.worldometers.info/coronavirus/country/china/

You will see that there trend is now a practically a horizontal line, which means the number of new cases per day is approaching 0. They are well beyonf the inflection point.

"If not, can we half ass predict when the IP will be reached?"
For that question I would say "defer to experts" as Eladar has already pointed out.

Here is why:
Otherwise in theory ( I believe ) it involves fitting a Logistic Curve to the current data in such a way as to minimize the sum of the squared error ( or some other method known method ) between the function and the data, then differentiate the function "f(t)" twice with respect to t ( t is our time variable ) and equate it 0. If you solved f"(t) = 0, for t you will determine the inflection point. Obviously this business could get fairly messy without some mathematical computational software.

@joe-shmo

I do not like the total cases because the US is doing a terrible job testing.

Deaths are more robust to testing because we do test sick people, especially those very sick.

So I would suggest what you are saying but with total deaths. It is a bit more morbid, but a bit more accurate.

The Ti 84 graphing calculator has a logistics regression option in the stat section. You go to the math column then down to option B.

@sonhouse said
@Metal-Brain
We could rule that out because C19 showed up in 2020 not in 19 in the US anyway.

You don't know that. No tests were available to prove that. You cannot rule it out.

The 19 in Covid 19 is for the year it appeared by the way.

https://www.globalresearch.ca/covid-19-originated-us-japanese-tv-broadcast/5704493?fbclid=IwAR0l2ZhW-A2zcMjJHERcYTZB36mjRvfgkIKJYRg6b7hAY35kQD2mBILhjBU

@Eladar
As of today the infections in US are 12,000 and 274 deaths.