KoVox Dataset

The KoVox Dataset provides structured data derived from promotional materials for 1,238 Korean classical vocal performances listed on NOL Ticket between 2016 and 2025. Each record represents a performance’s metadata and program structure extracted from web-based posters and brochures, which were converted from image-based text to structured form through OCR processing and MusicBrainz ID normalization. The dataset aims to preserve born-digital performance ephemera and to support longitudinal and comparative studies of Korea’s vocal music culture. Repository Contents Object Name Component Format Description KoVox_RDB Relational Database CSV (5 files) Structured tabular data representing the relationships among performances, works, persons, programs, and participations. KoVox_mhtml Web Archives MHTML Archived ticketing webpages following the “Golden Copy” principle, preserving original visual and structural layout. KoVox_ephemera Promotional Images JPG, GIF, PNG Posters and brochure images used as OCR sources; contain visual and textual design elements of performances.