U.S. Parliamentary Speech Corpus Processed for Analysis

Sv Corpora Parliament Processed is a collection of parliamentary speech data from the United States, processed for machine learning. The dataset, created by birgermoell, was last updated on Hugging Face in January 2022. It contains over 1 million text entries, as indicated by its platform size category.

Use Cases

• Train a BERT model for speaker role classification using the `speaker` and `text` columns.
• Perform sentiment analysis on political debates by applying NLP techniques to the speech `text` field.
• Conduct topic modeling to identify key legislative themes across sessions using the processed speech content.
• Analyze speech patterns and formality by calculating linguistic features from the `text` data.

Strengths

• Contains over 1 million text entries, providing a substantial corpus for analysis.
• Data is pre-processed and available in Parquet format for efficient handling.

Limitations

• Specific column names and data schema are not documented, complicating direct usage.
• The dataset's temporal coverage and update frequency within the source material are unknown.

Provenance

Source

Parliamentary proceedings from the United States (Region:us).

Collection Method

Speech data collected and processed from official records; specific methodology unknown.

Freshness

Last updated on the platform in January 2022.

Geography

United States