Thread by @AndrewJacksonV3

#CDCGate @DrDannielle
1)... Part 2 (09 May 2020). Last Saturday, I went looking on CDC's site looking for updates. As you may recall, I had an issue with New York State vs City. On a different page, I got confirmation that NYS does not include NYC.

Cont...

https://twitter.com/threadreaderapp/status/1258907689989615616

2)...What I can't find is the definitive death count due to #Covid_19? Anyhoo, decided to focus on Iowa, as it is a state w/republican governor & didn't have a stay at home order and reporting website is clear with an assortment of datapoints. As of 09May, 252 deaths.

Cont...

3)... From CDC, as mentioned in previous thread, the 08May update had duplicate entries for all datapoints. Upon closer look, 1 set is "as of 07May", 2nd is "as of 08May". Hmm, since I have a set for 06May, why not merge & see what we get? Jibes w/daily updates

Cont...

4)... CDC says that depending on States, it takes several weeks for numbers to stabilize. With Iowa, looks like 2 weeks... Week ending 18Apr remain unchanged, while week ending 25Apr & 02May (see #3 above) increases.

Cont...

5)... What else can we glean from CDC's "Provisional COVID-19 Death Counts by Week Ending Date and State", data.cdc.gov/NCHS/Provision… ?
A running count, week by week. The cumulative count, according CDC is 116/0/741 or 857 if all 3 columns are added. Arrgh. But IA says 252.

Cont...

6)...There's another CDC dataset, "Provisional COVID-19 Death Counts by Place of Death & State"(updated weekly) data.cdc.gov/NCHS/Provision…

8 categories for 'places of death'. Again, filtering only for Iowa. But these numbers don't match the prior image (updated daily, M-F)

Cont...

7)... That concludes my research from Saturday, 09May. What irks me is I don't know what CDC considers to be the definitive death count.
I'll monitor CDC for the next few days, waiting for updates.

Cont..

8)...BTW, here are 2 screenshots from Iowa's site, as of 10May. Reported deaths is 271--bearing in mind, CDC seems to be lagging by a week, as what they have is "As of 02May"
coronavirus.iowa.gov/pages/case-cou…

9...Waiting for CDC to update... @JackPosobiec flagged several states which had policies ordering nursing homes to accept infected COVID patients -- this was discussed elsewhere for past couple of weeks, but since I have the dataset, let's looksee.

https://twitter.com/MarkGudmunson1/status/1259715230298329089

Cont...

10)...Before we get into the above. Let's talk about definitive numbers, to whit; where are fake news getting their numbers?
Google as of 0936 on 11May2020 lists 80.6k deaths from COID-19, google.com/search?q=total…

But where did Google get their numbers?

Cont...

11)... Google got it from Wikipedia, eyeroll. en.wikipedia.org/wiki/Template:…

I've attached screenshots of their table template, footnotes and references.

So where did Wikipedia get their numbers from?

Cont...

12)...Wikipedia got it from, in part from CDC, cdc.gov/coronavirus/20… and from John Hopkins' CSSE Page, gisanddata.maps.arcgis.com/apps/opsdashbo… & USA Map Page, coronavirus.jhu.edu/us-map

Cont...

13)... I've discussed the CDC datasets above, but what about JHU's source?

CDC sez all states report to them eventually, but JHU apparently gets/scrapes their data from the individual States/Counties/Cities websites. No mention of reporting agencies providing it to them

Cont...

14)...If anyone can provide me the sauce to the definitive death counts, it would be much appreciated.

Back to the CDC's "by Places of Death" (POD) dataset. data.cdc.gov/NCHS/Provision…
As explained before, CDC's COVID death count leaves much to be desired.

Cont...

15)...I've tried several ways of adding/subtracting the columns to bring me close to 66k as reported by #FakeNews as of 02May. Looking at USA with 8 categories for POD.

To get % of deaths by places...
Column A = Columns (M-K) / I(Total Deaths)
Col B = Cols (M-K-H) / I

Cont...

16)... As you can see, the first manipulation is too low and the second is too high.
But let's look at the suspect Dem States (CA, MI, NJ, NY, PA) where COVID patients were dumped into Nursing Homes vs GOP states FL, GA, IA, ID, ND, TX

Negative % is garbage...

Cont...

17)...Back to drawing board, starting with USA only...

To get % of deaths by places...
Column A = Columns (H+(M-J-L)+K) / I (Total Deaths)
Col B = Cols (H+K+M) / I

As you can see from the scratch results on Row 444, both are way too high.

Cont...

18)... Using #17's construct; let's look at the suspect Dem States (CA, MI, NJ, NY, PA) where COVID patients were dumped into Nursing Homes vs GOP states FL, GA, IA, ID, ND, TX

Ratio-wise, the Dem states are high while GOP states are low.

Cont...

19)... The blogosphere has pointed out for awhile that #CDC is rolling up anything that may be COVID related, or even suspected as having #COVIDー19 COVID positive as a COVID death.
What if we ignore all underlying causes and focus only on USA COVID deaths as reported?

Cont...

20)...Using #19's construct; look at the suspect Dem States (CA, MI, NJ, NY, PA) where COVID patients were dumped into Nursing Homes vs GOP states FL, GA, IA, ID, ND, TX
Dem states are high while GOP states are low.
@realDonaldTrump @DonaldJTrumpJr @Scavino45
More to come later.

@threadreaderapp Unroll, thank you.

21)... I expanded #20 for all 54 States and Localities. Bear in mind, each is a % of the same line item.

Going further, I did an Excel SUMIFS function, to return a % of Nursing Home Deaths for that State's total COVID (8 categories each) Deaths.

I'll update all when CDC updates