• 14 Posts
  • 227 Comments
Joined 1 year ago
cake
Cake day: February 14th, 2024

help-circle

  • I don’t trust any data from a company that uses basis spline or hobby spline smoothing to represent data without showing the original points and a confidence interval.

    It’s an incredibly amateur move to display data this way. All the popular chart libraries include it because people like it, but objectively those people are wrong.

    I can rant for hours about this, but the smoothing on their charts actually shows values outside the real range from data. That’s okay if you use a smoothing model like loess with confidence intervals, but it is not okay as a “make it pretty” visualization technique.

    These spline smoothed data viz do not show any real trends between the points. The points can force the curve to make sharp jumps that aren’t real during sudden changes or changes in spacing of points. It’s more evident when the lines go below 0, but we don’t have that here.

    These charts do not replicate, as new data is added it actually changes the curve, which means you’re not plotting polling results over time anymore. The points are simply anchors on the curve, if you use illustrator for example draw a bunch of points, then hit the smooth button, that’s what’s happening here.

    These charts can have curves that go steeper than 90 degrees and actually loop backwards! We don’t live in the Jeremy Bearimy timeline, that can’t happen here. And it did with EKOS charts from a few months ago when they changed the time step of polling.



  • Just in case anyone is using 338 for info:

    • the higher voter turnout so far indicates this type of model will be less accurate this election.
    • the riding numbers are from provincial polling, there is no per riding polling incorporated. With increased turnout, I expect these to be less accurate.
    • the model works by rating pollsters based on how well the predict past elections. That’s why increased turnout means things may swing from their forecast.

    All that adds up to the potential for huge swings in votes, swings that can go either direction.

    Things to watch for:

    • Gen Z men’s rightward swing. Expect high turnout from this group for the Conservatives. They consume a lot of right wing media and have been struggling with tough economic conditions, so will be motivated.
    • Boomer influence is waning — this is the first election they are not the largest demographic.
    • Gen X are actually wealthier than boomers and are the only demographic that en masse had better economic outlooks since COVID. I don’t know if they’re going to stay the coarse with LPC or vote CPC for the tax cuts, traditionally more income swings right, especially with how Pollievre wants to change housing taxes
    • Older women and Quebec are very anti-Pollievre, this might end up with a suppressed vote or a very strong vote LPC.
    • Rural areas I think will swing more conservative than ever, and this might be where the forecasts swing. E.g. rural Ontario may not be as safe as thought.

    TLDR be prepared for surprises today.







  • I hope this happens and sends a message to the Conservative Party.

    You can’t be with the maple MAGA and the “anti woke” crowd while appealing to Canadians or even socially minded conservatives.

    You can’t only criticize and attack, at some point you have to demonstrate you can lead.

    When it comes to the protection of Canada, you don’t ever wait for the polls to tell you what to say. You drop everything and defend our country. Every time. The conservatives used to stand for that.

    Lastly: every single conservative candidate has declined to comment on news or show up to any local event in this country. What does that tell us voters about how accountable you will be in office?

    Anyways, get fucked Pollievre, I hope he loses his seat and the party splits, Canada deserves better than the CPC.





  • This is going to be much closer than polls and 338 suggest, and I base that on the historically high turnout.

    338 tunes polls to match elections. When you go from low turnout to historically high, that necessarily means people and demographics that previously did not vote are. That implies higher error than expected in the models, and a higher chance of an unanticipated result.

    Don’t rely on the polls to feel safe in the outcome, don’t rely on 338, get your butt out there and vote!

    (Side note, I am a Liberal voter, but this is my ABC advice)