Author Topic: COVID-19 Data Model Updated Daily  (Read 63109 times)

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 3224
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #195 on: April 10, 2020, 07:27:40 AM »
Is Stage-4 TDS considered an underlying morbidity?

 :o

Just can't escape it.  I had to look that one up.  Chucky Darwin will let us know. 

NEplay

  • Sunset Status
  • ****
  • Posts: 263
    • View Profile
    • Email
Re: COVID-19 Data Model Updated Daily
« Reply #196 on: April 10, 2020, 07:32:08 AM »
"Is Stage-4 TDS considered an underlying morbidity?"

I laughed out loud.

I have been following this thread and really appreciate the effort. Thank you. Most of it is above my pay grade but Beasho's numbers always seem to be days ahead of what the media eventually reports.

Is it to early to start getting data how many lives have been saved or not(?) by our sitting in our stew  instead of wafting through everybody elses?
New England Is My Playground

ninja tuna

  • Peahi Status
  • *****
  • Posts: 645
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #197 on: April 10, 2020, 08:49:42 AM »
Thanks for all this Beasho.

In terms of modelling. I have not heard any comparison to any other type of modeling.  Lots of moving parts, predicting into future time frames.   Being in Florida,  I see a good similarity to hurricane forecasting.  I have heard very little in terms any variation in models. Maybe I have missed it.  But it seems like I am only hearing about this IMHE model or whatever it is.  With hurricanes, they use quite a few models and form the infamous "cones of uncertainty".  Those usually go a week out and the hurricanes usually go outside of those 1 week prdecitions.

So someone comes up with a model with drastic numbers and every one jumps on it because "if it bleeds it leads" mantra of the news.  It is just like the hurricane that was coming towards Florida a few years ago.  Weather man comes on tv and tells everyone "IF YOU DON"T EVACUATE YOU ARE GONNA DIE!"    Lots of people stayed and there were under 5 deaths. Dont remember but very few.

Effing idiot

So basically what i am getting at is with all the weather modelling with all the their super-computing power missing the target so many times. Should we have expected the models to get this right?

And for it being blown out of proportion as much as it has been, imo, when something possibly more serious along the lines of an ebola type virus may rip around the world, a lot of people will not take it as seriously and there will be worse statistics. Just like the weather people say to prepare for all the hurricanes that never hit.  A lot of people got complacent for a while.

ninja tuna

  • Peahi Status
  • *****
  • Posts: 645
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #198 on: April 10, 2020, 09:51:45 AM »
and right after I post this someone in the Florida governors update says something along the lines of people in Florida getting complacent from hurricanes.

Then i found this.

https://chicagocitywire.com/stories/530092711-roseland-hospital-phlebotomist-30-of-those-tested-have-coronavirus-antibody

SUP Leave

  • Peahi Status
  • *****
  • Posts: 530
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #199 on: April 10, 2020, 10:21:58 AM »
Thanks Beasho.

Obesity is a true epidemic, but I guess the answer would be a "don't stay home" order forcing us out of our homes to the parks and beaches.

A lot of talk about models being incorrect, but all models are incorrect. If you think of them as a tool to elicit a response instead of a set of facts they make more sense. Beasho's models are more correct because he does not have any skin in the game in setting policy. If you work for the IMHE and you know that your model is going to be used to set policy, you are going to provide a wider range of outcomes, and use more conservative (less realistic) assumptions. No one wants their name on a CV-19 death.

The next question in my mind is: Don't we want to keep the infections level, now that the curve has flattened? I.e. throttling human interactions to keep the hospitals actively testing cases? I foresee testing data falling off soon, in that if we all stay home, and the hot spots are eliminated, no one will go get tested. Providing a false sense of the virus being eliminated and allowing normal life to resume and thus instantaneous flare ups? I mean we are going to be awash in PPE and ventilators, and hospitals are reducing staff. We will not have a vaccine for a year, so lets get on with herd immunity if we can support it with the hospitals current capability.

Just anecdotal, but one of the guys I coach youth sports with is a nurse. He is having his hours reduced 20% in order to avoid layoffs across the board at his hospital. I asked him about testing and he told me that they have plenty of tests but no customers. Our local county has 8 cases, but one is the same guy twice. He had a negative and then another positive a week later, so they counted it as new.

 

Make paddleboarding great again!

Tom

  • Cortez Bank Status
  • *****
  • Posts: 2993
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #200 on: April 10, 2020, 10:28:14 AM »
Thanks Beasho.

Obesity is a true epidemic, but I guess the answer would be a "don't stay home" order forcing us out of our homes to the parks and beaches.

A lot of talk about models being incorrect, but all models are incorrect. If you think of them as a tool to elicit a response instead of a set of facts they make more sense. Beasho's models are more correct because he does not have any skin in the game in setting policy. If you work for the IMHE and you know that your model is going to be used to set policy, you are going to provide a wider range of outcomes, and use more conservative (less realistic) assumptions. No one wants their name on a CV-19 death.

The next question in my mind is: Don't we want to keep the infections level, now that the curve has flattened? I.e. throttling human interactions to keep the hospitals actively testing cases? I foresee testing data falling off soon, in that if we all stay home, and the hot spots are eliminated, no one will go get tested. Providing a false sense of the virus being eliminated and allowing normal life to resume and thus instantaneous flare ups? I mean we are going to be awash in PPE and ventilators, and hospitals are reducing staff. We will not have a vaccine for a year, so lets get on with herd immunity if we can support it with the hospitals current capability.

Just anecdotal, but one of the guys I coach youth sports with is a nurse. He is having his hours reduced 20% in order to avoid layoffs across the board at his hospital. I asked him about testing and he told me that they have plenty of tests but no customers. Our local county has 8 cases, but one is the same guy twice. He had a negative and then another positive a week later, so they counted it as new.

Where do you live?

SUP Leave

  • Peahi Status
  • *****
  • Posts: 530
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #201 on: April 10, 2020, 10:50:46 AM »
Washington State

Grays Harbor County. Its a big county with not many people.

https://www.doh.wa.gov/Emergencies/Coronavirus

Make paddleboarding great again!

TallDude

  • Cortez Bank Status
  • *****
  • Posts: 5714
  • Capistrano Beach
    • View Profile
    • Email
Re: COVID-19 Data Model Updated Daily
« Reply #202 on: April 10, 2020, 11:50:04 AM »
Thanks Beasho.

Obesity is a true epidemic,
There may be a link to people who suffer from hypertension. Falling in that category are overweight people and people genetically predisposed to it. One of the blood pressure Rx type is an (ACE) inhibitor. The COVID-19 virus uses a receptor called ACE-2 to enter the cell. No conclusive tests yet linking the two, but there might be a connection.
It's not overhead to me!
8'8" L-41 ST and a whole pile of boards I rarely use.

digger71

  • Sunset Status
  • ****
  • Posts: 431
    • View Profile
    • Email
Re: COVID-19 Data Model Updated Daily
« Reply #203 on: April 10, 2020, 05:27:41 PM »

There may be a link to people who suffer from hypertension.

Found this for NY while searching for deaths by age group, but not sure if it exists on a national level.  Breaks out the comorbidities as well as age groups

https://covid19tracker.health.ny.gov/views/NYS-COVID19-Tracker/NYSDOHCOVID-19Tracker-Fatalities

Stew

  • Waikiki Status
  • *
  • Posts: 28
    • View Profile
    • Email
Re: COVID-19 Data Model Updated Daily
« Reply #204 on: April 11, 2020, 01:13:14 AM »
Live model here. 

I modified my 30 day death estimate to use exponential smoothing.  In other words you can't go EXPONENTIAL forever.  Brings down the 30 day estimate to 46K. 

New State Heatmap shown here.  What states are performing WELL as measure by Days to Double and Which are NOT performing well. 

NY, LA are emerging from the high burn phase.  Although NY now has more cases than any other SINGLE country (~ same as Spain).

https://app.powerbi.com/view?r=eyJrIjoiNzExYmM2ZWQtZWVmMS00OTI1LWE1NzYtMTU2OTk3ZGE4ZjMyIiwidCI6IjQzODE2ODE2LTY0OTYtNDU4My1iMmVmLWU5Y2U3MTg1NjI4MCIsImMiOjF9

Anyway to break Europe down into individual countries? Italy is on it's own. It would be interesting to see all the other countries split out as they're quite varied.

Admin

  • Administrator
  • Cortez Bank Status
  • *****
  • Posts: 6443
    • View Profile
    • StandUpZone
    • Email
Re: COVID-19 Data Model Updated Daily
« Reply #205 on: April 11, 2020, 02:50:56 AM »
If you work for the IMHE and you know that your model is going to be used to set policy, you are going to provide a wider range of outcomes, and use more conservative (less realistic) assumptions. No one wants their name on a CV-19 death.

Below is the note that IMHE posts with every model and every chart.  Any model that does not state similar is pulling your Willy.  An IMHE representative was also on the news last night and said that should strong distancing measures be relaxed in 2 weeks (which they consider far too early), their model (which was built on holding strong distancing measures until the end of May) would then show over 200,000 deaths by end of July.  Lastly, the latest IMHE model update showed the death count peaking two days ago but this has not yet occurred.  Were they incorrect?  Of course not.  You have seen the DIY models with sliders that others have posted here.  Tweak any variable by a tick and you have altered the results by 20,000 lives.  Lastly, there are nowhere near enough tests.  This is the coming fro the medical community.  That is not to say that there will not be surpluses in some regions.   

Uncertainty is the range of values that is likely to include the correct projected estimate for a given data category. Larger uncertainty intervals can result from limited data availability, small studies, and conflicting data, while smaller uncertainty intervals can result from extensive data availability, large studies, and data that are consistent across sources.

The model presented in this tool has a 95% uncertainty interval and is represented by the shaded area(s) on each chart.


« Last Edit: April 11, 2020, 03:36:36 AM by Admin »

surfcowboy

  • Cortez Bank Status
  • *****
  • Posts: 4929
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #206 on: April 11, 2020, 07:44:58 AM »
This has been great, but as was said, until we know how many people have this and don’t know it, we’re hosed.

Antibody tests are out there. LOS Angeles is testing 1,000 people a day now, randomly chosen, and anyone can apply. Over the next couple of weeks we will start to see those ghost cases that no one knew they had. We can make the call better then.

Articles (Stanford) now speculating that last Fall’s bad early “flu” in CA was C19. Long shot, but also would be another reason we were hit less hard.

I’m hoping for a 30% infection rate in CA right now as I type. It might not be there but a guy can hope.

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 3224
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #207 on: April 11, 2020, 08:55:11 AM »
The model presented in this tool has a 95% uncertainty interval and is represented by the shaded area(s) on each chart.[/b][/size]

This really pisses me off.  I work to make sure my reports are near 100% perfect.  There are a few small math issues with running totals and averages based upon how the data comes in but I would be willing to say my data is 98%+ accurate with regards to the source vs. 95% wrong.  Holy Moly.   

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 3224
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #208 on: April 11, 2020, 08:57:27 AM »

There may be a link to people who suffer from hypertension.

Found this for NY while searching for deaths by age group, but not sure if it exists on a national level.  Breaks out the comorbidities as well as age groups

https://covid19tracker.health.ny.gov/views/NYS-COVID19-Tracker/NYSDOHCOVID-19Tracker-Fatalities

More data on CoMorbidity and it is practically hidden.  I had to stitch together 3 Screen shots to put this view together.  I am not kidding you can NOT see this view anywhere.  You have to scroll through 3 rows at a time.

81% of Deaths in NY listed as CoMorbid.

Yet again numbers Closer to 100% than even 50% EVERYWHERE YOU TURN.  Obesity is NOT listed as a CoMorbidity.
« Last Edit: April 11, 2020, 09:02:37 AM by Beasho »

Beasho

  • Cortez Bank Status
  • *****
  • Posts: 3224
    • View Profile
Re: COVID-19 Data Model Updated Daily
« Reply #209 on: April 11, 2020, 09:01:35 AM »
Putting the data together from the website.  I could not scrape or find a programmatic source.   I had to type this data into a spreadsheet.

The CoMorbidities add to More than 200% of the deaths.  What!  This means the average person had 2.1 CoMorbidities. 

Add to this COVID-19 and you have 3.1X CoMorbidities. 

 


SimplePortal 2.3.7 © 2008-2024, SimplePortal