Turkish Parliament Speech Corpus Processed for NLP

Over 1 million text records contain processed transcripts from the Turkish parliament. The dataset was created by user 'mvip' and uploaded to Hugging Face in February 2022. It provides a corpus of parliamentary speech for natural language processing tasks.

Use Cases

• Train Turkish language models on formal political speech text.
• Analyze speech patterns and topic frequency across parliamentary sessions.
• Perform named entity recognition on mentions of people, organizations, and laws in the transcripts.
• Develop text classification models to categorize speech by political party or debate topic.

Strengths

• Contains over 1 million text records based on platform size tags.
• Data is pre-processed for NLP use, indicated by the title.

Limitations

• Specific column structure and preprocessing steps are unknown.
• Data is from a single country (Turkey), limiting geographic generalizability.
• Last update was in February 2022, potentially missing recent proceedings.

Provenance

Source

Turkish Parliament (inferred from title and 'Tr' prefix).

Collection Method

Transcripts processed from official parliamentary proceedings.

Freshness

Last updated 2022-02-23; update frequency unknown.

Geography

Turkey (inferred from 'Tr' prefix and 'Regionus' tag).