OpenWHO: Document-Level Parallel Corpus for Health Translation

OpenWHO is a document-level parallel corpus designed for machine translation evaluation in the health domain. It contains expert-authored, professionally translated educational materials for health workers sourced from the World Health Organization's multilingual e-learning platform. The dataset was created by raphaelmerx and was last updated on November 10, 2025.

Use Cases

• Train machine translation models for low-resource languages based on professionally translated health education content.
• Evaluate the performance of translation systems on domain-specific health texts.
• Analyze translation quality and consistency in expert-authored educational materials.
• Develop cross-lingual information retrieval systems for public health resources.

Strengths

• Contains expert-authored and professionally translated materials, suggesting high source quality.
• Specifically focuses on low-resource languages, addressing a gap in machine translation resources.
• Sourced from the World Health Organization's official e-learning platform, indicating authoritative content.

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.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

World Health Organization's OpenWHO.org e-learning platform.

Collection Method

Sourced from multilingual educational materials.

Freshness

Last updated 2025-11-10 10:44:58; freshness should be verified.