Dataset Card for Tamago Music Dataset

Dataset Summary

This dataset contains metadata for 1,567 music tracks from tamastream, a community-based music streaming platform built on the NEAR blockchain. The dataset includes detailed track metadata including titles, descriptions, genres, and user interactions, providing insights into independent artist communities and their music.

Languages

The dataset is monolingual:

English (en): All metadata including titles, descriptions and tags

Dataset Structure

Data Fields

The metadata for each track includes:

track_id: Unique identifier for the track (integer)
data: Object containing track details:
- track: Main track information object:
  - id: Track identifier (integer)
  - title: Title of the track (string)
  - description: Text description of the track content (string)
  - downloadDescription: Download related description (string or null)
  - duration: Length of the track in seconds (string)
  - downloadPrice: Price for download (null or number)
  - downloadCount: Number of downloads (integer)
  - imageUrl: URL of the track cover image (string)
  - originalImageUrl: Original image identifier (string)
  - playCount: Number of plays (integer)
  - userId: Creator's user ID (integer)
  - genreId: Genre identifier (integer)
  - displayOrder: Display ordering (null or integer)
  - downloadFilename: Filename for downloads (string or null)
  - downloadContentType: Content type for downloads (string or null)
  - trackStatus: Status of the track (string)
  - token: Track token (string or null)
  - tokenUserId: Token user identifier (integer or null)
  - listing: Listing information (null or object)
  - createdAt: Creation timestamp (string)
  - updatedAt: Last update timestamp (string)
  - User: User object containing:
    - id: User identifier (integer)
    - username: Username (string)
    - nearAccount: NEAR blockchain account (string)
    - firstName: First name (string)
    - lastName: Last name (string)
    - websiteUrl: Website URL (string)
    - avatarUrl: Avatar image URL (string)
    - userRole: User role type (string)
    - accountStatus: Account status (string)
    - lastLoginAt: Last login timestamp (string)
    - verifiedAt: Verification timestamp (string or null)
    - createdAt: Account creation timestamp (string)
  - Genre: Genre object containing:
    - id: Genre identifier (integer)
    - name: Genre name (string)
    - parentId: Parent genre ID (integer or null)
    - iconId: Icon identifier (integer)
    - createdAt: Creation timestamp (string)
    - updatedAt: Last update timestamp (string)
  - TagsForTrack: Array of tag objects:
    - id: Tag identifier (integer)
    - name: Tag name (string)
    - createdAt: Tag creation timestamp (string)
    - TrackTags: Track-tag relationship object
  - Reactions: Array of reaction objects:
    - userId: User identifier (integer)
    - trackId: Track identifier (integer)
    - commentId: Comment identifier (string)
    - referenceId: Reference identifier (string)
    - referenceData: Reference data (null or object)
    - reactionType: Type of reaction (string)
    - trackAt: Track timestamp (string or null)
    - createdAt: Reaction creation timestamp (string)
    - User: Reactor's user information object
  - Comments: Array of comment objects

Data Splits

The dataset is organized into three splits by file type:

Split	Number of Instances
metadata	1,567
images	1,567
audios	1,567

The metadata split contains JSONL files with track information and metadata
The images split contains cover artwork images for each track
The audios split contains the music track audio files

Each split corresponds to the same 1,567 tracks but separates the different file types for easier handling.