19th-Century Spanish Newspaper OCR and Page Images from the National Library of Spain

Spanish publications from the Biblioteca Nacional de España's Hemeroteca Digital provide over 830,000 high-resolution page images and full-text OCR. The dataset contains roughly 800 million text tokens from approximately 40,000 documents primarily from the 19th century. Author ferjorosa uploaded this collection to Hugging Face, with a last recorded update in 2026.

Use Cases

• Train or evaluate historical OCR models based on the high-resolution page images.
• Analyze linguistic trends in 19th-century Spanish based on the full-text OCR content.
• Conduct digital humanities research on publication topics and frequency based on the document collection.
• Fine-tune language models for archaic or historical Spanish text based on the 800 million token corpus.

Strengths

• Large scale with over 830,000 pages and approximately 40,000 documents.
• Includes both machine-readable text (roughly 800 million tokens) and high-resolution source images.
• Sourced from the authoritative Biblioteca Nacional de España (BNE) - Hemeroteca Digital.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Temporal coverage is primarily 19th century but includes titles extending into the early 20th, which may introduce period inconsistency.
• Description metadata is limited; actual data quality and OCR accuracy require manual inspection after download.

Provenance

Source

Biblioteca Nacional de España (BNE) - Hemeroteca Digital

Collection Method

Sourced from digitized historical publications, with OCR processing applied.

Time Range

Primarily 19th century, with some titles extending into the early 20th century.

Freshness

Last updated 2026-01-09 12:56:46; freshness should be verified.

Geography

Spain