2
BadBoyCornPop 2 points ago +2 / -0

You are doing gods work.

I too saw the "drop and roll" in my plots.

There's some nuance to TGP numbers, in their columns incremental_vote_pct_dem/rep.

Nail that in a reproducible way and I think we overturn all these states.

Super behind on work this week - PUMPED.

1
BadBoyCornPop 1 point ago +1 / -0

This file is of Oakland only but seems to have the same sort of data

https://results.enr.clarityelections.com//MI/Oakland/105840/269402/reports/detailxls.zip

There may be similar files on the same website

This is a simplified view of the Oakland data:

http://www.climateaudit.info/data/election/MI_Oakland-2020.csv

and here's my attempt at sorting by "PC1", https://thedonald.win/p/11Q8lS215l/x/c/1BkBNOLxps

Edit: see also this thread https://twitter.com/ClimateAudit/status/1326905529160126464

3
BadBoyCornPop 3 points ago +3 / -0

I need help now

I have no time to look at this in the next few days.

I have been looking at Dr. Shiva's plots and McIntyre's similar plot here:

https://twitter.com/ClimateAudit/status/1326905529160126464

and have done something similar in C#, but using a proprietary IDE (sorry)

Code is here:

https://pastebin.com/TUWevCuC

spez: data here: www.climateaudit.info/data/election/MI_Oakland-2020.csv

data derived from data here: https://results.enr.clarityelections.com//MI/Oakland/105840/269402/reports/detailxls.zip

What I think I have done:

  1. Read McIntyre's csv file
  2. Created a PCA (Principal Component Analysis) using the columns MixElectionD, MixElectionR, StrElectionD, StrElectionR, divided by TotVotesTotal
  3. Added PC1-4 to each record
  4. Ordered the results by PC1
  5. Plotted the 4 columns a) raw b) scaled by TotVotesTotal

These are the resulting plots (similar to those from Shiva/McIntyre);

https://i.maga.host/yFXlJpj.png https://i.maga.host/mpZiLe7.png

PC1 is, roughly, how democrat or republican a precinct has voted.

The relationships for StrElectionD, StrElectionR, MixElectionR fit intuition and are roughly in same ordering of PC1.

MixElectionD makes no sense. This shows no visible relationship with how democrat or republican the other votes are.

I need help to know:

  • have I messed up in implementation? (I do not know stats and have never used the library used to compute PC1)

  • have I messed up in interpretation?

  • how would this look with other data?

  • wtf is MixElectionR/MixElectionD anyway

  • MixElectionD - Are these Biden only ballots?

Has anyone else been looking at this?

Spez2:

I have also plotted the results ordered by PC2, which seems to be, roughly, MixElectionD / TotVotesTotal.

https://i.maga.host/q8O2VkA.png

Hope I'm not fooling myself - based stat pede is having a look now - thanks!

1
BadBoyCornPop 1 point ago +1 / -0

Another problem is that a simple application of Benford's law does not work well where the county sizes are not randomly distributed. For example, if many counties have 1000 seats it then it's unlikely that Trump or Biden would have a vote number starting with a 1. Digits 3-6 would be more likely.

by Doctor
1
BadBoyCornPop 1 point ago +1 / -0

deleted.

1
BadBoyCornPop 1 point ago +1 / -0

Mellissa Carone (Dominion Whistleblower)

14:00 - "Supervisor's name is Nick and he's also co-owner of the company, Dominion voting."

spez:

Nick Ikonomakis - Vice President-Development at Dominion Voting Systems Corp.

Mellissa says that it's insider information that he is also a co-owner.

1
BadBoyCornPop 1 point ago +1 / -0

Bill Kennard appears in position 45 in Obama's ambassador auctions

https://twitter.com/JunkScience/status/775890679104937984

3
BadBoyCornPop 3 points ago +3 / -0

Externally embedding the video is disabled.

nsfwyoutube link: https://www.nsfwyoutube.com/watch?v=4p8Aq5BM9io

view more: ‹ Prev