Dataset Viewer
The dataset viewer is not available for this dataset.
Unexpected token '<', "<html> <h"... is not valid JSON

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

DFS Glossary — Amharic & Afaan Oromoo

Expert-verified glossaries of Digital Financial Services (DFS) terminology in Amharic (am) and Afaan Oromoo (om), published as structured, machine-readable, openly-licensed data.

Open language infrastructure for two low-resource Ethiopian languages — for developers, researchers, translators, and the financial-inclusion community.

Languages Amharic (am, Ge'ez script) · Afaan Oromoo (om, Latin script)
Entries 87 Amharic + 86 Afaan Oromoo = 173
License CC BY 4.0
Version 1.0

Usage

from datasets import load_dataset

am = load_dataset("shega/dfs-glossary", "am")   # Amharic glossary
om = load_dataset("shega/dfs-glossary", "om")   # Afaan Oromoo glossary

print(am["train"][0]["term_local"])  # should print clean Ge'ez, e.g. የአገልግሎት ነጥብ

Data fields

Field Type Description
id string Stable, language-prefixed id (am-001, om-001). Never reused.
term_en string English term — the join key across languages.
term_local string Term in the target language (Ge'ez / Latin).
language string am or om.
category string | null Domain grouping (e.g. Channel, Concept, Process). Verbatim from source.
definition_local string Verified definition in the target language.
definition_en string | null English definition, where available.
example_local string | null Usage example in the target language.
example_en string | null English usage example, where available.
version string Dataset version (1.0).

The English term (term_en) is shared across the two languages, so the am and om configs can be joined on it.

Methodology

Translated and verified by domain experts as part of the AKOFADA project, and built under a strict fidelity policy:

  • Source is the editable copy decks, never the PDFs. Ge'ez (Amharic) script corrupts when copied out of PDFs; for a terminology dataset a single wrong character destroys trust. PDFs are used only for human spot-checks.
  • Text is preserved verbatim — only whitespace tidying is applied. No spelling, wording, casing, or Unicode normalisation of the verified text.
  • The two languages are cross-linked on the English term.

Provenance

  • Producer: Shega
  • Project: AKOFADA
  • Funding: Gates Foundation

Citation

@dataset{shega_dfs_glossary_2026,
  title  = {DFS Glossary — Amharic & Afaan Oromoo},
  author = {Shega and AKOFADA},
  year   = {2026},
  version = {1.0},
  publisher = {Shega},
  note   = {DOI pending — to be added on first release},
  url    = {https://huggingface.co/datasets/shega/dfs-glossary}
}

License

Creative Commons Attribution 4.0 International (CC BY 4.0). Free to share and adapt, including commercially, with attribution.

Downloads last month
25