Data Analysis of Tweets by Senators on Recent Issues Facing Americans¶
An NLP-driven analysis of 419,000+ tweets from 100 U.S. senators, revealing that congressional communication is overwhelmingly shaped by party affiliation rather than constituent concerns.
Abstract¶
Tweets are expressions of feelings on a topic; they show what a person is thinking, what they value, and how they communicate. Twitter is being increasingly used by politicians to connect with their constituents. We analyze the last six months of tweets of 100 senators of the 117th Congress. We show what issues they talk about, what issues matter to them, and how frequently they discuss them. We show that although senators are supposed to represent their entire state, they mostly adhere to party-line politics and only talk about issues related to their party, failing to represent the entirety of topics relevant to Americans in the current period.
Key Contributions¶
- Collected and analyzed 419,000+ tweets from 100 U.S. senators over six months (Jan--Jun 2022), stored in a SQLite database for structured querying
- Mapped tweets to 35 key issues identified by the Washington Post using NLTK WordNet synonym expansion, growing the issue lexicon to 100+ terms for robust matching
- Demonstrated clear partisan patterns: Republican senators emphasize gas, prices, inflation, economy, and cost, while Democratic senators prioritize climate, economy, opportunity, prices, and school
- Generated word cloud visualizations and issue frequency distributions per senator, showing clear ideological alignment in communication patterns
- Found that senators largely fail to address the full spectrum of issues facing their constituents, instead communicating along party lines regardless of state-level concerns
Links¶
- PDF: to be hosted
- arXiv / TechRxiv: to be added