Download
The dataset is data, not a document — take it in whatever form fits.
YAML — full dataset ↗
untded/untded-2005 — nine category files covering every element, each with its source page and extraction confidence, plus the extraction and verification code (Ruby, lutaml-model).
CSV & JSON exports ↗
Release artifacts: elements.csv (one row per element) and elements.json (full records), straight from the dataset.
JSON-LD
/data/untded.jsonld — the whole directory as a linked-data graph: every element an identified resource with schema.org and UNTDED vocabulary terms.
RDF Turtle
/data/untded.ttl — the same graph as Turtle, for SPARQL stores and RDF tooling. The vocabulary is defined in the dataset’s YAML-LD context.
Compact index
/data/index.json — a light JSON list of all 1504 elements: tag, name, representation, change tag, status, category.
Source PDFs & provenance ↗
untded/references — both PDFs, the front-matter OCR, the printed legend, and the D.05B cross-check corpus.
Attribution
Content: UNTDED 2005, ECE/TRADE/362 (in parallel ISO 7372:2005), © United Nations / UNECE. Digitized by this project on behalf of UN/CEFACT and ISO/TC 154; reproduced with attribution. Code: MIT. When you use the dataset, cite the source publication and this digitization.