This algorithm accurately predicted November 2016 voter turnout using only public data

An algorithm developed in 2015 tries to guess 2016 election results.

My first campaign role was as a volunteer canvasser.  A few of my college friends and I went door-to-door for one of our professors who was running for congress.  He ended up winning.

I quickly learned a basic rule of campaigning: campaigns have limited funds, and they have limited time.  It in no way is a marathon.  It is a mad sprint– a very strategic one where you only have so much fuel.

 

‘Who’ you talk to can be more important than ‘how many’ you talk to.

Was your audience made up of all registered voters?  Was your audience likely to be persuadable, and even more important, were they likely to vote?

Maybe just a few particular voters are the reliable voice that shapes the direction of an entire precinct.

This isn’t a poll, and it isn’t an aggregation of a bunch of polls.

This is a fully scalable dynamic algorithm that analyzes data specific for each voter.

So far I am pleased with the data results.

Imagine the possibilities.

With even more here: dankleinman.org/tag/elections/

dan@dankleinman.org