Using Data Scraping and NLP to Understand Former Trump Voters

I wrote a Python script that scraped over 1,000 hours of former Trump voters on YouTube, converted those audios into texts, ran a sentiment analysis of those texts using Natural Language Processing, and then clustered those texts into overlapping themes. The goal was to develop a theoretical model to understand why they voted for Biden.

In our current political climate, dismissing those whose political leaning differs from ours is easy. On the surface, it may appear that we have nothing in common. There were a growing number of former Trump voters who publically shared why they decided to switch party. However, nobody could understand why.

I wrote a Python script that scraped over 1,000 hours of former Trump voters on YouTube, converted those audios into texts, ran a sentiment analysis of those texts using Natural Language Processing, and then clustered those texts into overlapping themes. The goal was to develop a theoretical model to understand why they voted for Biden.

There were a few big takeaways from my analysis. First, former Trump supporters were invested in America. Second, they were driven less by political issues (covid, the economy, taxes) but more by hope, values, and traditions. Finally, many saw themselves, their families, and their past leaders as contributors to America’s greatness through their small and large sacrifices. Some talked about joining the army, serving in politics, or being active in the political process.

The results highlighted the importance of talking about values over interests. Former Trump voters were more concerned with the values of Romney, McCain, Bush, Reagan, and Biden. They cared about American institutions and wanted America to succeed, even if it was under Biden.

Ultimately, they felt Trump was distinctly un-American because he didn't have any of those values. They saw him as selfish and greedy. While they disagreed with many of Biden's policies, they felt Biden shared their values of institutions, decency, respect, and sacrifice for the greater American community.

My experience with the voice of customer research has taught me that the "why" is 1000 times more potent than the "what." If I could tap into their "why's", then perhaps I could better connect with them about their whats. 

In the end, it’s not the “what” that matters. It’s the “why” that counts because in between the “what” and the “why” are points of connection. The insights from this project enabled me to identify core areas of interest that serve as points of connection for someone who wishes to connect with Trump voters.