US Presidential debates have been contentious in the past, but never to the extent that the first debate in the lead-up to the 2020 Presidential Election was! Both President Trump and former Vice President Biden generate a lot of emotion amongst both their supporters and their detractors. Wouldn’t it be cool if we could track the emotional response of people to the debate as it progresses? With Receptiviti, now we can.
Using LifeRaft’s social media mining platform we scraped tweets containing the keyword #presidentialdebate from 1 hour before the debate started until 30 mins after it finished. Tweets from outside the United States were discarded, leaving 3,127 tweets. Of these, 1,148 tweets explicitly mention Trump but not Biden, with 670 mentioning Biden but not Trump. We grouped all the tweets into 5 minute bins and by which candidate they were targeted at as well as the physical location of the tweeter (by latitude and longitude, to within the nearest 0.05 degree). After some minor cleaning (e.g. removal of URLs) we then ran all the aggregated tweet samples through the Receptiviti API. The resulting data set provides a geographically tagged timeline of US Twitter users’ emotional responses to the first US Presidential debate at a resolution of 5 minutes. To visualise this, we built an app using the Shiny library in the R programming language to compare the geographic spread and intensity of emotions across the continental US throughout the debate.
You can find the app here:
We’ll leave it to you to play with the app and come up with your own analysis. Suffice to say, there’s a lot of anger and disgust directed at both candidates throughout, but also a lot of love and joy. Unsurprisingly the one thing we see very little of is boredom, as love it or hate it the debate definitely kept everyone engaged!