Govdocs1: 220,000 Open-Access Government PDFs for OCR

Approximately 220,000 open-access PDF documents, comprising about 6.6 million pages, from the govdocs1 dataset. The dataset was uploaded by BEE-spoke-data and last updated on December 29, 2025. It is intended to be processed with optical character recognition (OCR).

Use Cases

• Training and evaluating optical character recognition (OCR) models based on the raw PDF content.
• Analyzing document structure and layout in government publications based on the PDF format.
• Building search indices or information extraction pipelines based on the textual content after OCR.
• Studying language patterns and topics in public sector documents based on the corpus.

Strengths

• Large scale with approximately 220,000 documents and 6.6 million pages.
• Includes a randomly sampled subset of 5,000 PDFs for immediate preview and testing.
• Open-access nature facilitates use without restrictive licensing barriers.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

govdocs1 dataset

Collection Method

Collection of open-access government PDFs, likely from public websites.

Time Range

null

Freshness

Last updated 2025-12-29 04:40:27; freshness should be verified.

Geography

null