A sentiment-based political debate scoreboard

The Democratic Party presidential debates just kicked off. With so many candidates (enough for two nights of back-to-back debates), I wanted to build something to gauge widespread political perception in realtime. To do this, I decided applied VADER Sentiment Analysis to a live feed of tweets from the Twitter API. The results were fascinating.

I fed an average of the VADER compound polarity scores of all tweets associated with solely one candidate directly into the scoreboard. The compound score outputs a continuous value ranging from +1 (most positive) to -1 (most negative).

Check out this segment where Kamala Harris attacks Joe Biden on race and segregation. Note how Kamala’s score climbs significantly, and Joe’s score drops.

Or this segment, where Pete Buttigieg’s score rise as he talks about racial profiling and police brutality.

It’s important to note that other factors could be at play here – in particular, bots that may be included in the data feed. Moreover, this doesn’t take into account retweets and likes due to the real-time nature of the feed – a future version of this scoreboard would likely incorporate those as well.

Regardless, it’s interesting to see how public perception changes in realtime on such a large scale!

Check out the code to this project here. To actually run the scoreboard, you’ll have to apply for a Twitter API Key.