Philly will bring it home

EDay voter turnout reports from Philadelphia Great!

And I don't have to rely on long anecdotes. To illustrate the point numerically, please look at the Sixty-Six Wards voter turnout tracker. It's a very cool app where people can specify their ward and ward (what normal cities call a district), when they voted, and what their voter count is. The algorithm then estimates the total amount for the entire city. If you are a fellow resident of the City of Brotherly Love, be sure to log into the app and be more specific.

I also went one step further. I looked at historical (and apparent) patterns of strong turnout in the morning and evening and a small dip at lunch and built a model from that.

Context: 749,000 total votes in 2020

Status now:
1. 185,000 mail-in ballots were returned.

2. The last update was at 10:10 a.m., with the tracker estimating 216,000 personal EDay votes.

3. (This is where my model will improve throughout the day): 10:10 am is 24% of the voting window, which I estimate is about 39% of the total votes to be cast.

4. By extrapolation, we get an estimated total (Mail+Eday) of 853,000, 14% more than in 2020.

If that number is anywhere near correct, we've broken the voting record in Philly, and we're all but assured of a big win in PA (and hopefully nationally).

Update (10:45 a.m. data): To date, 282,000 have been personally estimated. The projection for the final is 835,000, +11% compared to 2020.

Update (11:17 a.m. data): This is literally unbelievable (and to be clear, it's possible that some of this is noise or malicious actors entering fake data – although there are safeguards in place for this). Expect it to jump around a lot. An estimated 350,000 in person. The extrapolation for the final is 945,000, +26%. That's a number so good that I think both the 66 precincts and my own model for when I expect votes to be cast must be flawed. But that means we'll get closer to a truer value as the day goes on.

Update (1:07): I think we ruined it. Sixty-six Wards has not been updated since 11:45am. At this point, the projected total vote count was 905,000, up about 21% from 2020. I should note that I adjusted things down a bit to only include 157,000 mail-ins because I think that's all there is was previously accepted (hence the difference from the previous ones). Anyway, I wrote to the developer, who presumably already knows about the issue, but hasn't received a response yet.