Author Topic: COVID-19 Data Model Updated Daily  (Read 19230 times)

Admin

  • Administrator
  • Cortez Bank Status
  • *****
  • Posts: 5242
    • View Profile
    • StandUpZone
    • Email
Re: COVID-19 Data Model Updated Daily
« Reply #165 on: April 05, 2020, 12:25:08 PM »
Hi Beasho,

Cuomo from yesterday:

Gov. Andrew Cuomo said on Saturday that New York had about seven days to prepare for the coronavirus apex in the state, warning that “we’re not yet ready for the high point.”
“Our reading of the projections is: We are somewhere in the seven-day range,” Cumo said at a press briefing on the COVID-19 crisis. “Four, five, six, seven, eight-day range.”

As I mentioned, when New York reaches ~5000 cases the state will pass the Country of Italy in deaths/population.  That will happen in the next few days. 


surfinJ

  • Teahupoo Status
  • ******
  • Posts: 1741
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #166 on: April 05, 2020, 02:39:07 PM »
Colorful graphs and dry numbers.  Any empathy here?  “One for all and all for one”? 

My buddies and family in NY are describing a very grim ambiance.

Tracking with Germany?  Are you just in the numbers and missing the drama being experienced by your countrymen? 
Germany’s non partisan built up and extremely robust healthcare system is the reason. They were ready.

As far as the data available, I know from my friends experience that they are not testing or accounting.
And now I have read that there is significant under reporting of virus deaths due to the lack of testing.

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 2570
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #167 on: April 05, 2020, 02:39:25 PM »
The numbers still don't add up.  The United States is NOT ITALY.  NY - Maybe but Unlikely.  THIS IS GOOD NEWS!

Let's even say that NY goes the way of Italy.  (This is a tall order because there is NO SINGLE municipality in the United States that is close to the Italy's death%).   Italy has a 12.3% death rate.  NY stands at 2.8%.  We have to go back 5 weeks to find Italy at a 2.8% death rate.  At the beginning of their outbreak.  They then had 1 Day Death % of 20% to 25% of New CASES Every Day for 2 Weeks.

Eventually the Total % Death Rate climbed to 12.3% of Cases. 

Let's assume NY will climb.  NY is still just one State.  One state with <<10% of the population.  Look at the 2nd chart.  There is another 160MM people in States with Half as many current deaths.  10X the population with slightly more than half the current deaths.  The other states don't have the density of people and behavior has changed. 

There is no way the rest of the country catches up.
« Last Edit: April 05, 2020, 02:41:28 PM by Beasho »

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 2570
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #168 on: April 05, 2020, 03:01:47 PM »
Colorful graphs and dry numbers.  Any empathy here?  “One for all and all for one”? 
. . . I know from my friends experience that they are not testing or accounting.
And now I have read that there is significant under reporting of virus deaths due to the lack of testing.

If you know it then share the source.  If Germany is so much better than the United States and we are doomed to a gruesome future then share the evidence.   

I find comfort in these numbers and if I can help a few others contextualize what is happening then great.  Without the numbers we are at the whim of Snake Oil salesmen and Shaman Hand-wavers.  Fox vs. CNN.  Without numbers there is NO SCIENCE. 

I am reminded of a quote from a fellow Ithacan:

“At the heart of science is an essential balance between two seemingly contradictory attitudes--an openness to new ideas, no matter how bizarre or counterintuitive they may be, and the most ruthless skeptical scrutiny of all ideas, old and new. This is how deep truths are winnowed from deep nonsense.”  CARL SAGAN
« Last Edit: April 05, 2020, 03:05:38 PM by Beasho »

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 2570
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #169 on: April 05, 2020, 03:12:28 PM »
Scientists are providing a wide range of models for the US as a whole depending largely on our performance as a society, but most have us at over 100,000 deaths even with a nationwide (all states in compliance) stay at home order.  You are basing your assumptions largely on start dates that are believed to be incorrect and testing data that is known to be inaccurate.

Here is another Dry and Callous perspective.  But I do think this guy knows how to build a model:

https://www.cnbc.com/2020/04/05/bill-gates-coronavirus-pandemic-a-nightmare-scenario.html

« Last Edit: April 05, 2020, 03:22:42 PM by Beasho »

Admin

  • Administrator
  • Cortez Bank Status
  • *****
  • Posts: 5242
    • View Profile
    • StandUpZone
    • Email
Re: COVID-19 Data Model Updated Daily
« Reply #170 on: April 05, 2020, 03:16:11 PM »
The numbers still don't add up.  The United States is NOT ITALY.  NY - Maybe but Unlikely.  THIS IS GOOD NEWS!

Hi Beasho,

Your own estimate from the last page of 82,000 deaths for the USA is 5.2 times the raw death count of Italy's current death count of 16,000.  The death/population would be equal using those figures.  The figures in your last post would make the US significantly worse.

In what way is that not worse than Italy?


Beasho

  • Cortez Bank Status
  • *****
  • Posts: 2570
    • View Profile

PonoBill

  • Cortez Bank Status
  • *****
  • Posts: 23249
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #172 on: April 05, 2020, 06:28:00 PM »
I did some sloppy modeling, looking at all the published models and data sets for the USA. I start with the notion that we have no idea of how many people are infected and could be wrong about when the big curve started. Move the inception date back a week, assume 10X for the real infection rate, and we're at the peak. Actually, we kind of were there yesterday.

This is no more than a guess, but we should see some useful data in the next few days. I think the read denominator is more like 20X, but I have no evidence for that, just a hunch. I keep looking at the death curve and I think it's more a measure of how wadded up the hospitals are than a measure of how bad the virus is. Anywhere that's under stress it's a lot worse, and it doesn't seem to take much stress to kick up the death rate. Looking at the data from here: https://covidtracking.com and here: https://covid19.healthdata.org/projections  it appears that any time the hospitals get close to full ICUs the death rate climbs quickly. 

I should mention that flattening the curve moves the peak out into the future. I guess that might be obvious to people used to reducing data, but it was a little bit of a surprise to me.
« Last Edit: April 05, 2020, 07:11:52 PM by PonoBill »
Foote 10'4X34", SIC 17.5 V1 hollow and an EPS one in Hood River. Foote 9'0" x 31", L41 8'8", 18' Speedboard, etc. etc.

stoneaxe

  • Cortez Bank Status
  • *****
  • Posts: 12057
    • View Profile
    • Cape Cod Bay Challenge
Re: COVID-19 Data Model Updated Daily
« Reply #173 on: April 05, 2020, 11:23:41 PM »
Surprised?....isn't that the point of flattening the curve? Attenuation stretches it all out. The end will be even further into the future too.
Bob

8-4 Vec, 9-0 SouthCounty, 9-8 Starboard, 10-4 Foote Triton, 10-6 C4, 12-6 Starboard, 14-0 Vec (babysitting the 18-0 Speedboard) Ke Nalu Molokai, Ke Nalu Maliko, Ke Nalu Wiki Ke Nalu Konihi

dietlin

  • Sunset Status
  • ****
  • Posts: 251
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #174 on: April 06, 2020, 05:40:43 AM »
The crux of the matter:


it appears that any time the hospitals get close to full ICUs the death rate climbs quickly. 

Admin

  • Administrator
  • Cortez Bank Status
  • *****
  • Posts: 5242
    • View Profile
    • StandUpZone
    • Email
Re: COVID-19 Data Model Updated Daily
« Reply #175 on: April 06, 2020, 06:54:02 AM »
We are seeing tops in daily confirmed deaths in many countries and it would be great if the US joined them in the next few days (or has joined them).  These counties are seeing massive losses at the top of their curves and the nature of the downslope from this first upsurge is as of yet unknown.  This graph current from this morning at https://ourworldindata.org/coronavirus is super useful because it shows the curves beginning when each country reached 5 deaths.  That lines things up pretty closely so you can get a better sense of relative escalation over days.  It uses a rolling seven day average to round out potential daily oddities.  This is a very good for judging how well the curve has been flattened in each location. 



« Last Edit: April 06, 2020, 07:29:43 AM by Admin »

SUP Leave

  • Sunset Status
  • ****
  • Posts: 477
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #176 on: April 06, 2020, 09:03:32 AM »
Thanks again Beasho for the data and standing strong with it.

I am like you, I feel much better looking at the data myself. Empathy has nothing to do with it. Knowing makes me comfortable.

Early in this epidemic, I believed China was being honest in their data collection and dissemination and it gave me a false hope. They provided harmful data to the rest of the world. Their numbers were never adding up even from the earliest cases in American cities. Anytime the media states "studies from China" in any report it should be taken with an ocean of salt.

Of course cities/states/media in the US are great at feeding out catastrophe. It is an addicting endorphin for some people. So much of it is just theater IMO.

WA (my state) has hit the peak in deaths and hospital needs for this round. WA and OR are returning over 400 ventilators, and the prior week Governor Inslee was demanding Trump start the Defense Protection Act to make more ventilators.

It will be really interesting to see how the roll back into "normal" works, and how the CV cases respond. WA governor Inslee is planning on wearing the COVID success like a badge as he makes another run for president, and my inclination is that he will keep on the strict stay-home order longer than most just to prove it.
Make paddleboarding great again!

SUP Leave

  • Sunset Status
  • ****
  • Posts: 477
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #177 on: April 06, 2020, 11:08:14 AM »
BEASHO,

IMHE (UW) (CDC) just revised their national death toll do 81,766 -  so your 82k number is tracking perfect right now, but by end of day on 4/9 you are going to end up being too high.
Make paddleboarding great again!

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 2570
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #178 on: April 06, 2020, 02:08:25 PM »
I saw a briefing with Andrew Cuomo last night saying that with full measures adhered to his numbers are showing 16,000 projected deaths for NY state.  Population of 20 M. (800 deaths per million).   He also said that he believed that the actual number which Dr Fauci had given the president was 94,000 deaths nationally if all measures were strictly adhered to nationally (all states).  With a population of 330 Million in the US that is 284 deaths per million people. 

In business we call what the president did on Sunday 3/29, announcing 1.1MM to 2.2MM deaths as possible "establishing a baseline."  This way you have an objective to measure against and succeed.

He, or Fuaci, then warned of 100K to 240K possible deaths given the measures taken. 

The president would NOT let that number get stated unless he could provide MORE successful results.  Therefore YES the real number is likely below 100K.
I suspect the real number could be HALF of that ~ 50,000.
BUT if it were 94K and it resulted in 284 deaths per 1MM that would be just above the data I showed for 2017 mortality of 17.1 Deaths per 100K from Influenza or 171 Deaths per Million.

I had to go find what my latest estimate was and I put down 50,000.  Yes I see that HealthData.ORG revised down to 81,766. This is pretty good news.  My ballpark was / is still on the low end of their estimates.

82,000 was for Illustrative purposes. let's say the high end of my range. 

http://www.healthdata.org/covid/updates

However the 82,000 was part of an even BOLDER prediction which would be ZERO NET DEATHS For 2020. 

When taken over the entire year the Corona Virus would just displace 82,000 other Chronic Disease @ Risk people.  The net result for 2020 would be ZERO incremental total deaths from Corona Virus.  With the reduction in travel and general @ risk behavior the Net number may be negative.
« Last Edit: April 06, 2020, 02:20:39 PM by Beasho »

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 2570
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #179 on: April 06, 2020, 03:09:32 PM »
We are seeing tops in daily confirmed deaths in many countries and it would be great if the US joined them in the next few days (or has joined them). . .

I have looked at the trailing Death count but just built it into a view using Trailing 5 Days.  The US is JUST about there.