As a data scientist myself, this sort of analysis warms my heart. But of course I stumbled upon this right before going to bed, so couldn't resist doing a little bit of exploration! Will be pouring through this a lot more tomorrow, but wanted to at least take a glance at Wisconsin.
What is very puzzling is that, if I'm interpreting the data correctly, at the time of that last "push" by the dems to nudge the D/R ratio above 1.0, the GOP had a commanding lead. From 2:58 AM to 9:37 AM (again, assuming I'm reading the data correctly), with 262 batches of votes, the minimum ratio was 0.905, max was 0.989, and average was 0.924. Translation: GOP held onto a consistent advantage throughout this time span as various magnitudes of batches came in.
The GLARING issue is the one large batch of votes that followed that was 85.1% dem, which just so happened to be enough to erase the GOP lead, and actually put the GOP several thousand votes behind. The following batches comprising >30k votes maintained a constant ratio: 49.3% incremental for dems against 49.0% for GOP. Strange.
One plausible explanation is that this one burst was from one of the urban areas. Given that it's the largest one it might be reasonable to conclude it was Milwaukee. But if you look at the top 10 batches received, the average dem percentage was 49.8%. Top 5 was 53.7%. Top three: 64.6%. So it seems like this county really bucks the trend! And then if this were truly Milwaukee, hey maybe they just have a history of voting this dramatically? Not even close. From 2000 to 2016 the dem candidate percentages were: 60.7, 62.3, 68.2 (Obama), 68.2 (Obama again), 65.6. So that 85% figure...awfully suspicious. You might wonder: what if this wasn't Milwaukee? Wouldn't make a difference. If you looked at any of the other large counties that's about as blue as they get.
To recap, to believe this is pure and true one would have to believe:
It was coincidence that this was just enough to erase a huge lead
It was coincidence that the largest reported batch broke far and away the hardest for the dems
That the voters were so enchanted with Biden that he received 19 points higher as a voting percentage than Obama did in either election
My friends, if this isn't fraud I would LOVE to read a rational explanation.
Take a look also at vote share columns. Only 3 decimal places. Should be 7 to prevent vote shaving when stat reduced, like from 0.493 to 0.492.
Depending on total votes this kack of precision can swing votes around greatly.
This is causing votes to disappear, why trump looses thousands in batch updates.
Good way to shave votes if this really is the max precission being recorded and sent around.
Edit: appears csv file has this issue, checking out json file instead.
As a data scientist myself, this sort of analysis warms my heart. But of course I stumbled upon this right before going to bed, so couldn't resist doing a little bit of exploration! Will be pouring through this a lot more tomorrow, but wanted to at least take a glance at Wisconsin.
What is very puzzling is that, if I'm interpreting the data correctly, at the time of that last "push" by the dems to nudge the D/R ratio above 1.0, the GOP had a commanding lead. From 2:58 AM to 9:37 AM (again, assuming I'm reading the data correctly), with 262 batches of votes, the minimum ratio was 0.905, max was 0.989, and average was 0.924. Translation: GOP held onto a consistent advantage throughout this time span as various magnitudes of batches came in.
The GLARING issue is the one large batch of votes that followed that was 85.1% dem, which just so happened to be enough to erase the GOP lead, and actually put the GOP several thousand votes behind. The following batches comprising >30k votes maintained a constant ratio: 49.3% incremental for dems against 49.0% for GOP. Strange.
One plausible explanation is that this one burst was from one of the urban areas. Given that it's the largest one it might be reasonable to conclude it was Milwaukee. But if you look at the top 10 batches received, the average dem percentage was 49.8%. Top 5 was 53.7%. Top three: 64.6%. So it seems like this county really bucks the trend! And then if this were truly Milwaukee, hey maybe they just have a history of voting this dramatically? Not even close. From 2000 to 2016 the dem candidate percentages were: 60.7, 62.3, 68.2 (Obama), 68.2 (Obama again), 65.6. So that 85% figure...awfully suspicious. You might wonder: what if this wasn't Milwaukee? Wouldn't make a difference. If you looked at any of the other large counties that's about as blue as they get.
To recap, to believe this is pure and true one would have to believe:
My friends, if this isn't fraud I would LOVE to read a rational explanation.
Take a look also at vote share columns. Only 3 decimal places. Should be 7 to prevent vote shaving when stat reduced, like from 0.493 to 0.492.
Depending on total votes this kack of precision can swing votes around greatly. This is causing votes to disappear, why trump looses thousands in batch updates.
Good way to shave votes if this really is the max precission being recorded and sent around.
Edit: appears csv file has this issue, checking out json file instead.