Data Analysis: Pennsylvania Database Still Missing Over A Million Voter Voting Records
Every state in the United States has a more or less transparent approach to elections and share much of the raw data with the public. Pennsylvania is no exception. In a state with many complaints of campaign fraud, voter data is updated weekly and can be downloaded for $20 per download.
I downloaded four dataset snapshots of Pennsylvania's voter rolls updated on November 2, 9, 16, and 23 respectively to compare and analyze to see if the fraud complaints were reflected in the database.
The comparison revealed that, as of November 20, the number of voters in the Pennsylvania database who had participated in this poll differed by more than one million from the total number of ballots published in the media.
Voter Database Information Should be Consistent with The Total Number of Voters
Let's start with some basic numbers: the database table contains over nine million rows of registered voters, and each voter profile contains 153 columns, including voter ID, name, address, party, school district, precinct, method used in each election, Date Last Changed, and so on. The first we should look for is the cumulative voter total, which should be available in every state's election database, but may be expressed in different ways.
For example, the Nevada voter registration dataset is cut off in the evening of November 6, and when you see a column in the database with the attribute "Election Day", counting how many "2020-11-03" there are in that column yields 1,394,715. With 1,405,376 voters on the Nov. 15 news, the two numbers in Nevada basically match.
In another example, Washington State allows mail-in ballots, and the database I downloaded was cut off on November 2, and by the same token, we can add up the "2020-11-03" in the "Last Voted" column. That results in 3,230,733, compared to the 3,545,289 early voters reported in the news. In other words, as of November 2, 3.23 million voters had cast ballots in the Washington State database, which is close to the 3.54 million voters reported in the news, with a one-tenth of a percent difference probably due to the difference between the morning and evening cutoff times, which was within normal range.
It is common sense that the total number of voters reported by the state or news organizations must be roughly the same as the total counted in the database, and that is the point of the following discussion.
The Pennsylvania database has been missing information on millions of people who voted.
In the Pennsylvania database, out of 153 attributes, the column Last Vote Date = '2020-11-03' for attribute #26 has a similar content. This is shown in the following figure.
As you can see, the cumulative number of votes cast over the past three weeks has been increasing: 2,404,653 on November 1st, 3,722,445 on November 8th, 4,472,083 on November 15th, and 5,850,725 on November 20th.
The problem is that on November 15th, for example, the number of votes cast in Pennsylvania government and news media was 6,857,562, including 3,423,976 for Biden, 3,354,836 for Trump, and 78,750 for Jo Jorgenson. In contrast, the database showed only 4,472,083 people have voted, a difference of 2,385,479. As shown in the chart below.
In other words, as of November 15, more than 2,385,000 ballots had appeared from unknown sources, and more than a third of them were not recorded in the ballot database at all. This corroborates the Pennsylvania phenomenon described by the Trump campaign's Director of Data and Strategy Matt Braynard's team on November 16: they surveyed 1,137 people after the election, and more than a third of those who mailed in their ballots could not find a record of having voted on the government's verified web page.
By November 20, almost all of the votes should have been processed, and the gap was still about a million.
What does it mean when the total number of votes in the database is consistently and seriously lower than the real number of votes?
The Pitfalls of Duplicate Voting
This means that the ability to update the "Last Vote Day" in the database, to reject duplicate votes, is selectively disabled; or that the news reports of the votes cast are fictitious, with millions of fictitious ballots.
Here is a brief explanation of the importance of this "update vote date" feature in the database.
Normally, if A voted, the system scanned in, and the database was updated some label that he had voted. A went out for a spin and came back in to vote again, and the voting system checked, "Oh, his voting record has been updated to '2020-11-03' and he's already voted." He wouldn't be able to vote.
If the "Last Vote Day" feature in the database for a third of the population failed, they would be able to repeat voting multiple times and the system would keep accepting it! This explains the strange phenomenon that on one side, 1137 people, more than a third of whom sent their ballots, could not find in the database that they had voted (the date of voting was not updated), while on the other side, the state government reported over a million more ballots than were in the database.
It is technically easy to specify a certain period of time, a certain polling place, and certain voters as not having to "update their Last Vote Day," but this is discriminatory treatment of voters, which is a violation of election laws. It also allows fraudulent ballots to be swept in multiple times without being rejected by the system! It can be said that this is a channel for illegal ballots to mix with legal ballots.
Of course, the author is doing an analysis of the direction of the database and its functionality, where technical errors should be taken into account. For example, a database crash, program malfunction, improper operation, etc. All technical errors could cause data disorder and loss, and the senior database administrator should immediately notify the voter to suspend the election, repair it, and then repeat the election. However, we did not find any such notification. So we have reason to suspect that many of these millions of extra ballots were of unknown origin.
The "Mail-in Ballot Dataset" Has Been Tampered With
In light of the fact that the State of Pennsylvania has tampered with the "mail-in ballot data set" made public in the Epochtimes after the article "Pennsylvania ballot data suspicious, 220-year-old man votes by mail" was published and the public is no longer allowed to download the mail-in ballot open data, it is likely that the numerical gap described in this article will be fixed in a week or two.
Thus, my question to the State of Pennsylvania and its senior database administrators is: What mechanism did you use in the month before and after the election to ensure that there were no duplicate ballots mixed in? This is an eternal question, not one that can be explained by tampering data!
Assuming a recount, a separate scanner must be used to record the voter ID on all recounted ballots, and as soon as duplicate ballots appear, they should be invalidated. In the end, I would not be surprised to find that the number of legitimate ballots is significantly less than the published figure of 6.9 million. Because that is how absurd the Pennsylvania database is trending right now.
I strongly encourage data analyst/database professionals to study the databases of states with election problems and find out what the facts are and what the problems are. Download and save your evidence.
For $20 each, the official Pennsylvania registered voter dataset:
https://www.pavoterservices.pa.gov/Pages/PurchasePAFULLVoterExport.aspx
Translated from Chinese Epochtimes: https://www.epochtimes.com/gb/20/11/24/n12572869.htm
Comments
Post a Comment