Hugging Face's logo

Datasets:
SmartQHSE
/
hse-qa-corpus

Tasks:
Question Answering
Text Generation
Text Retrieval
Languages:
English
Size:
n<1K
Tags:
question-answering
instruction-tuning
rag
hse
ehs
occupational-safety
License:
Dataset card Files
xet
 Community
hse-qa-corpus / README.md
SmartQHSE's picture
SmartQHSE
Add Related SmartQHSE datasets cross-link section
6b4d84c verified
preview code
|
raw
history blame contribute delete
7.16 kB
metadata 
license: cc-by-4.0
language:
  - en
tags:
  - question-answering
  - instruction-tuning
  - rag
  - hse
  - ehs
  - occupational-safety
  - osha
  - iso-45001
  - construction-safety
  - oil-gas-safety
  - gcc-regulations
  - long-form-qa
pretty_name: SmartQHSE HSE Q&A Corpus
size_categories:
  - n<1K
task_categories:
  - question-answering
  - text-generation
  - text-retrieval

SmartQHSE HSE Q&A Corpus

DOI Wikidata

24 long-form HSE / occupational-safety question-answer pairs across 15 categories — incident rates, ISO 45001, permits, risk assessment, OSHA (US), HSE (UK), GCC regulations, PPE, heat stress, exposure, ergonomics, incident investigation, training, HSE software. Each answer is multi-paragraph with cited sources, formulas, and OSHA/regulatory references. Suitable for instruction tuning and RAG.

Citation (preferred — academic) SmartQHSE Ltd (2026). SmartQHSE HSE Q&A Corpus [dataset]. Zenodo. https://doi.org/10.5281/zenodo.20010252

DOI: 10.5281/zenodo.20010252
Zenodo record: https://zenodo.org/record/20010252 Wikidata entity: Q139623115 CC BY 4.0 — free to use commercially with attribution.

Files

File Description
data.jsonl One JSON record per line — primary format. Loadable via datasets.load_dataset() directly.
data.json Same data as a single JSON array (with array key matching the source).
data.csv Selected fields as CSV for spreadsheet users.

Live access (CC BY 4.0, CORS-open, no auth)

Loading example 

from datasets import load_dataset
ds = load_dataset("SmartQHSE/hse-qa-corpus")
print(ds["train"][0])

# Or directly via the live REST API
curl https://www.smartqhse.com/api/v1/answers

License

This dataset is released under Creative Commons Attribution 4.0 International (CC BY 4.0). Free to use commercially with attribution. Cite as:

SmartQHSE Ltd (2026). SmartQHSE HSE Q&A Corpus [dataset]. https://www.smartqhse.com/answers

Considerations for using the data

  • Small dataset (24 entries as of v1). Will expand quarterly.
  • Geographic bias toward US OSHA, UAE OSHAD-SF, UK HSE, GCC regulations. Less coverage of Asia-Pacific, Africa, Latin America.
  • Industry bias toward construction, oil & gas, manufacturing — reflects SmartQHSE customer mix.
  • Answers are point-in-time (last_updated field). Regulations change; cross-check with the canonical_url before relying on a specific number.
  • English-only at v1. Arabic translation planned for v2.

Related datasets

Sources

  • US Bureau of Labor Statistics (BLS) — SOII + CFOI
  • US Department of Labor — OSHA Injury Tracking Application + standards
  • UK Health and Safety Executive (HSE)
  • European Commission + Eurostat ESAW
  • IOGP, ILO, NIOSH, ACGIH
  • UAE OSHAD-SF, KSA SAPI, Qatar QCDD, Oman MOLSD
  • ISO Technical Committees (TC 283 — OH&S Management Systems)

All consolidated and republished under CC BY 4.0 with attribution.

About SmartQHSE

SmartQHSE is the AI-native HSE/QHSE platform for construction, oil & gas, manufacturing, and industrial teams. We publish open data because the broader HSE profession deserves free access to the safety statistics our trade bodies otherwise gate behind expensive memberships.

Related SmartQHSE datasets

Part of the SmartQHSE open HSE data collection (catalog · data paper DOI 10.5281/zenodo.20010657 · HuggingFace org):

Companion full-text corpus: /llms-full.txt — 213KB markdown reference compiled for LLM ingestion. All datasets CC BY 4.0.