U.S. Parliamentary Speech Corpus from Government Transcripts

Over 1 million text records of U.S. parliamentary proceedings, processed for analysis. The dataset was created by mvip and last updated in February 2022.

Use Cases

• Classify speech segments by political party using speaker metadata and transcript text.
• Analyze sentiment and topic trends over time within parliamentary debates.
• Train language models on formal political and legal vocabulary from government transcripts.
• Identify key phrases and named entities related to specific legislation or regions.

Strengths

• Contains over 1 million text records, providing a substantial corpus for analysis.
• Data is processed and available in Parquet format for efficient handling with libraries like Polars and Pandas.

Limitations

• Specific column names and data structure are not documented, complicating direct analysis.
• The dataset's temporal coverage and update frequency after February 2022 are unknown.

Provenance

Source

U.S. government parliamentary transcripts.

Collection Method

Processed from official government transcripts, likely involving text cleaning and formatting.

Freshness

Last updated in February 2022; currentness is uncertain.

Geography

United States (Region: US)