3,065,408 articles and approximately 2.5 billion tokens extracted from historic British newspapers published between 1800 and 1896. The collection originates from the Heritage Made Digital program at the British Library and consists of text generated via Optical Character Recognition (OCR).
Use Cases
- Train language models on 19th-century English using the 2.5 billion tokens of OCR text
- Perform historical trend analysis across the 1800-1896 timeframe using article publication dates
- Conduct topic modeling on Victorian-era journalism using the 3,065,408 article-level text entries
Strengths
- 3,065,408 individual articles extracted from digitised newspaper pages
- Covers a 96-year historical period from 1800 to 1896
- Includes approximately 2.5 billion tokens of OCR-processed text
- Sourced from the British Library's Heritage Made Digital digitisation program