NewTypeCola has published the VOKU Archive on Hugging Face, a collection of 66,010 scanned pages from student-run campus radio broadcast cue sheets in South Korea. The documents span from 1988 to 2019 and include OCR text, pseudonym-token annotations, and metadata.
- The dataset preserves 32 years of writing from a university broadcasting station.
- It includes page images, processed text, and structured metadata.
- A companion GitHub repository details the AI-assisted processing workflow and human review steps.
The project aims to keep these historical documents readable and accessible beyond their original storage.