Silkworm - US presidential election forecaster

I’ve just updated and re-released a US presidential election forecaster. This uses poll aggregation to forecast the result of the 2020 US presidential election. Of course, it’s written using Bokeh.

Here are some screen grabs:

Geography1373×784 139 KB

PollViewer1371×498 59 KB

State1371×785 90.2 KB

VotesDistribution1366×779 58.4 KB

You can get the project here: GitHub - MikeWoodward/Silkworm: Silkworm - US presidential election predictor using opinion poll data. Uses Python, Pandas, Bokeh.

I’m giving a talk about this project at PyData Boston. Given the fact that this is COVID time, my talk will be given over Zoom, so it doesn’t matter where you are, you can still attend.

My talk will cover:

• Real polling experience
• What went wrong in 2016
• Election forecasting methods
• Poll aggregation methods
• Silkworm demo
• How Silkworm was built (code generation, Bokeh) and some lessons learned
• Lessons from the whole project

You can sign up for my talk here: https://www.meetup.com/PyData-Boston-Cambridge/events/272845963/

Hi @Mike_Woodward,

Thanks so much for sharing this! Is it okay if we tweet about it? And if so, is there a Twitter handle we should tag?

It’s great work, and Im sure it will generate a lot of interest!

–Carolyn

Hi Carolyn,

You can tweet away!

I don’t have a Twitter handle I’m afraid.

It would be nice if you could include a link to my talk: https://www.meetup.com/PyData-Boston-Cambridge/events/272845963/ - but if that’s against the rules, or it’s too much, I would be happy with just the tweet!

Thank you!

Mike