Election Data Analyzer

Let's make it happen, nobody is coming to save us.

In case you are not aware, you can sign up for emails of press releases from some of your SoS. I have signed up for a bunch just to see what is being discussed. The CO SoS email is typically a self-congratulatory pat on the back at how great they are in maintaining a "gold standard". Please.....what propaganda. How do I know? I have looked in detail, down to the individual voter ID per address level to discover how much TRASH is in the registration roles. Data integrity? No way. I will gladly share the specifics with the Clerks if they would listen. How about the case in Weld County where 400+ people are registered at 3 duplex apartments because they are attending a 3 week CDL class in that town and are crashing at these apartments while they are in town? How about the case of 100's of people registered at a hotel in Mesa County near a CDL school? Or the dozens of homes in Broomfield County with 10 or people per residence registered to vote with different names? These are the facts but they don't make it easy to find it, you have to really dig. Registration role data integrity is one of the cornerstones of one legal vote per person and these roles are a mess. I have been told the same thing about other states. Maybe I should just route the press releases to spam? What the SoS want you to believe is if they COUNT the votes thoroughly, we have a good system. Ridiculous. People are not that uniformed anymore. That horse is out the barn for good. Vote Amish sounds good to me.

EDA v4.0 will be released on Monday through USEIP. There is a call scheduled and posted now on the EDA team. I am not going to give those details here. You need to join USEIP. USEIP members will have access to the new version after that call.

Hey folks, I want to address some technical topics I get asked about. EDA process up to 1MM records on a single tab. Why is that? It is an Excel limitation. The actual Excel limitation is 1048576, so I just cut it off at 1MM. This accommodates around 70% of all the counties in the US in one analysis. EDA is meant to be used at the precinct, city, district and county level. It is not a statewide or multi-state analysis tool. You need a database for that. EDA will always be a spreadsheet not a program for a variety of reasons which I will not enumerate here. The most important one is DECENTRALIZATION. You don't need an internet connection to use it and the data is saved on your machine not in the cloud. In terms of processing times. I have one user who can process 1MM EDA records in about 10 seconds. He is running Linux as his OS with Windows10. I like to test worst case when developing software, so, for that, I am using my 10+ year old Mac running Windows on a virtual PC with Parallels. 1MM records takes me just a few minutes to analyze (while I warm my coffee). The point is, whether it is 10 seconds or 3 minutes is irrelevant. The bulk of your time is going to be spent analyzing the static table output, looking at maps from the export file data and canvassing. This will take hours, days and weeks. Finally, once I run an analysis, I never clear it. I just start with a new blank file, paste in the new data and run it again. There are wide variety of computers in the user base (over over 500 people now) with different OS, RAM and processing capabilities. But, you do need Excel 2016 or later. I am not going to support backward compatibility beyond that version of Excel which is now 2 releases behind the latest one. I also will never provide an unlocked copy of EDA to anyone. There are a wide variety of reasons for this. Let's focus on ELECTION INTEGRITY case studies and analysis here. Thank you.

Hey folks, this is off topic, but I am going to mention it anyway. I don't go off topic much....EDA is being provided as a public service and that will never change. I don't charge for it and I do not accept donations. It is distributed through USEIP and I am proud to be a part of that volunteer organization. I have put 100's of hours into EDA in terms of product development, training, questions, case studies, etc...I enjoy the work and it is where I can contribute and make a difference in Election Integrity. Now on to the "off topic" part of this thread. Given the wide spread usage of EDA and positive feedback, I am considering "hanging up my own shingle" as a consultant who "takes bad spreadsheets and turns them into professional spreadsheets" as a monetized service. If this has any interest to your business, please contact me at "electiondataanalyzer@protonmail.com". We can discuss your project needs further. Now back to the regularly scheduled program..... I am about 90% done with EDA updates and am on track to release next weekend. Cheers.

Hey folks, wrapping up a major update to EDA and will issue v3.9 for release exclusively through USEIP. If you want a copy of EDA v3.9, read the FAQ. Updates will include more user configurable text and number properties, additional pivot tables for those new properties, and birth date to age conversion calculation, phone number and address compare analysis tables and improved walk list functionality. I am hoping to have development wrapped up next week along with all the training instructions updates. It's a lot of work...Targeting next weekend for a release on USEIP with a training call. There also will be some social media exposure after the release with an extended interview about EDA v3.9 (more specifics on that later after it is scheduled). We have 100s of users on USEIP working together on election integrity and EDA is one component of that. Come on over and join the party. After I get through this release, I will start posting specific case study examples again as I have done in the past. Cheers.

Hello folks. I have had some personal matters keep me away from my EDA work. I am trying to ramp back into it over the next week or so. This channel has been quiet for a while, just thought I would put that out there. I am surveying our USEIP members to see what other useful features I might add to EDA. If you want to post something here as a new feature request, please do. We will see what the most popular ideas are in about 7-10 days. Cheers.

Hey folks. I am going to focused on EDA enhancements for awhile. Going to pause on the case study posts until after v3.9 is ready. FYI.

The meeting for tonight listed through USEIP EDA Team has been changed to 7pm EST. I had a conflict come up. If you are a member of USEIP EDA Team, you are welcome to attend. General Q&A. I am not going to post the meeting link here. Thanks.

I deleted the red barn post. I made a mistake when copying the address. Sorry for the misdirection.

Some of you may have been on the excellent call put on by the FL team about canvassing last night. I want to address one point made on that call. The state of FL has very high quality voter role data that is readily available. There are many states out there that do not have nearly that good of a data set available. I am working with one state now that has very sparse data. For example, they do list list registration date of the voter. Also, the methodology described on that call for the most part was dependent on using a data base or a web application to manage your team. Hopefully all of you will progress to that point for your state. EDA is not a database solution. Therefore it cannot use the methodology for finding ghost voters and ghost registrations that was defined on that call. That methodology is a "Bobby Piton" method that requires using a database and crunching millions of records across an entire state. It is a centralized solution. EDA is a decentralized solution. There are others ways to find these ghosts. One of them for example is unique address counts of registrants. EDA does provide this. I know it works. There are many other ways to find ghosts once you become familiar with the EDA spreadsheet. EDA is one component of a comprehensive state canvassing program. You can get started and be productive with EDA first but always drive to incorporate it's use into a more sophisticated model with a state analytic lead.

Another update on my new initiative. I reran the EDA with combined APT#'s on the address to get rid of the false positives (so to speak) on registration counts. The new "winner" is this beautiful barn home where the mapping application says 458 people are registered. I checked with the state lead and this town is bad news so that jives...but, as with everything else, I can't trust the map 100% so the locals are going to check it out. I did double check the source file and that is not the issue. 458 people do indeed have the EXACT same address, not APT #...to be continued.

Hey folks. I recently got involved in a new initiative. I requested the canvass team put there data through EDA without the APT # on the address and then again with the APT# on the address. WITHOUT the APT# on the address, the top unique address had 1063 people registered against it. With the APT#, the same address does not show up in the top 100 unique address. When I plotted this one address without APT# on the map, the location showed a strip mall. Now, either they scraped the mall and put up a large apartment apartment building or there is something very nefarious going on here. This requires a canvass which they will do soon. There is BIG lesson learned here, if the canvass shows fraud.....the bad guys hide ghost registrations by adding unique APT# onto addresses. It effectively will mask the count of registrants by address calculation in EDA.....To Be Continued....You may want to consider running your address through EDA again IF you have a data set that separates the APT# from the street address. Good luck.

I looked at 9 USPS locations in a single county and found 27 registrations. 20 of them are Active and 17 Voted. The RESIDENTIAL address in the registration database includes the street address of the post office but it also includes a "Apt" number or "Unit" or number, not a PO Box (not that it matters except to indicate perhaps they are trying to hide something)? I am going to dig further on this and make sure there are not any special cases in this state that would permit such a thing....but, I thought you would find this interesting to think about. This took me about 30 minutes to locate the USPS address and then I used the Pivot Filters tab in EDA to find the registrants. I exported those records and developed these pivot tables.

PO Box Case Study: Before I commit to a big dig, I try and take small sample size and see what comes of it. Maybe there is a way to make the dig more efficient? Maybe it is a dry hole not worth pursuing? Well, in one county I found 29 locations where mail is delivered between 8 different companies. There were a total of 12 voters using address at those 29 locations. I would say what I have found so far is an indication I should continue to look more. I am going to try an urban county next and see what happens. This took me about 2 hours using the EDA tool and an on line map you are all familiar with.