An overly simplified way to track candidate polarity(Better AI version next).
Virginians ONLY!
Here is the mean average polarity of pro-Youngkin tweets vs pro-McAulliffe tweets.

Super Oversimplified:
This graph is simply averaging the number of pro-candidateX tweets. Each tweet gets +1. These ~1,600 tweets were from self-proclaimed VA residents and had very specific polarized hashtags like #winwithglenn(pro-Youngkin) or #teamterry(pro-McAuliffe). Note that this is very oversimplified and simply used to contrast vs the nationalized version.
Everyone's Opinion!
Here is a similar graph of ~33,000 polarized(hashtagged) tweets from anywhere in the world. You can see inflection points around 10/18 and then at 10/29.

NEXT UP! Using AI to classify un-hashtagged tweets as pro-Youngkin or pro-McAuliffe
Comments