Turkish Podcast Collection with Audio and Transcripts

A collection of Turkish-language podcast episodes, likely containing both audio files and associated text transcripts. The dataset was created by the user 'Aynursusuz' and was last updated on Hugging Face in March 2026. Platform tags suggest it contains at least 100,000 items and may be stored in Parquet format.

Use Cases

• Train Turkish automatic speech recognition models using the audio files and corresponding text transcripts.
• Analyze podcast content and topics by processing the text transcripts for named entities or sentiment.
• Develop audio-text retrieval systems by linking the audio modality with the text modality for search tasks.
• Perform speaker diarization or identification on the audio files across multiple podcast episodes.

Strengths

• Contains at least 100,000 items based on the 'Size Categories100 Kn1 M' tag.
• Includes both audio and text modalities, enabling multimodal machine learning tasks.

Limitations

• Specific row count, column details, and audio duration statistics are unknown.
• The geographic and demographic representativeness of the podcast speakers is unclear, potentially introducing bias.

Provenance

Source

Hugging Face, user Aynursusuz.

Collection Method

Likely scraped or aggregated from various Turkish podcast sources.

Time Range

null

Freshness

Last updated March 2026.

Geography

Primarily Turkey or Turkish-language content.