YouTube Music Links and Artist Metadata for Discovery

Over 12 million links to music on YouTube, compiled by LAION. The dataset was created by recursively exploring the 'Fans might also like' artist graph, extracting metadata such as artist names and subscriber counts. It was last updated in November 2024.

Use Cases

• Train graph neural networks on the 'Fans might also like' connections to model artist similarity.
• Analyze the correlation between an artist's subscriber count and their position in the discovery graph.
• Build a content-based music recommender using artist metadata and associated song lists.
• Study the growth and connectivity of music genres by mapping the discovered artist network.

Strengths

• Contains 12 million music links, providing a substantial base for analysis.
• Includes artist metadata like name and subscriber count for contextual analysis.
• Built via recursive graph exploration, capturing a wide network of related artists.

Limitations

• The exact number of rows, columns, and sample data are unknown, limiting reproducibility.
• Potential geographic bias as the initial seed list and YouTube's 'related artist' algorithm may favor certain regions.
• Data freshness depends on the crawl date; artist subscriber counts and song catalogs may be outdated.

Provenance

Source

YouTube, via the Hugging Face platform.

Collection Method

Recursive graph exploration starting from a seed artist list, using YouTube's 'Fans might also like' feature.

Freshness

Last updated on 2024-11-14.

Geography

Likely global, but specific coverage is unknown.