There’s a job opening for the President of the United States and citizens are fortunate to have a vote in one of the world’s most important job interviews. During the first presidential debate, our Data Science team used the HireVue Insights deep learning engine to take a look “inside the interview”- analyzing the video, audio and language patterns of the presidential candidates.
This week, with the second Presidential debate behind us, we analyzed how the candidates’ sentiment and emotional intelligence (EI) evolved from the previous debate, as well as how each individual compared to other.
First, let’s take a look at how Donald Trump’s emotion and temperament evolved from the first debate (black), to the second (red). You can see less surprise, but significantly more sadness, anger, disgust and negative valence.
Now let’s take a look at Hillary Clinton’s emotional intelligence, which also evolved from the first debate (black) to the second (red). You can see more expressions of smiles and joy, and significantly less expressions of fear and surprise.
Next, we have a comparative of both candidates EI during the second debate. This is the overall sentiment from the entire debate, including all topics, questions and interactions. Clinton is in blue, Trump is in red.
Just for fun, our team also analyzed the audio, video and language from Gary Johnson’s recent town hall. While the questions and setting were different, in general, he shows more surprise than Clinton, but less than her first debate; and much less sadness, anger and disgust compared to Trump. He was the lowest on negative sentiment, and split Trump and Clinton on expressions of smirk, joy and smiles.
Below are the word clouds for both candidates which is based on frequency of terms. Trump world cloud:
Clinton word cloud:
Stay tuned for more data from our data science team in the next debates. Learn more about HireVue Insights and how Hilton Worldwide cut time to hire from 6 weeks to 6 days.
Disclaimer: we don’t have a dog in this race, nor is this part of the normal HireVue Insights data model which correlates such attributes to actual job performance data.
This is just our data science team’s idea of a fun public service. The analysis uses emotion sensing analytics of the candidates’ audio, video and language patterns.