NLP
Hausa Text Dataset
Short Hausa sentences for language identification, classification and literacy tools.
Dataset description
A demonstration corpus of everyday Hausa: greetings, market talk, weather and household phrases. It is fictional sample text written for this product preview, not a field collection and not licensed for commercial training.
- Creator
- AI HUB Data Team (demo)
- Records
- 100,000
- Format
- JSONL
- Collection period
- Demo catalogue · not field-collected
- Updated
- 2026-11-07
- Downloads of samples
- 0
Quality score 96/100
Research sample
Samples use a 10-minute grant. Full files are not hosted.
Quality breakdown
20% completeness + 18% accuracy + 14% consistency + 12% annotation agreement + 14% documentation + 12% duplicate health + 10% metadata.
Completeness
98%
Accuracy
95%
Consistency
94%
Duplicate rate
1.2%
Annotation agreement
93%
Documentation
96%
Metadata
92%
Data preview
Safe sample only. Restricted fields are not shown.
| id | text | language | domain |
|---|---|---|---|
| h1 | Ina kwana, yaya kake? | hau | greeting |
| h2 | Ina zuwa kasuwa yau. | hau | market |
| h3 | Ana ruwa a Kano. | hau | weather |
Data card
- Purpose
- Language identification and Hausa text classification benchmarks.
- Collection method
- Authored demo sentences. No contributors, no scraped pages.
- Annotation method
- Each row is tagged language=hau by the catalogue author.
- Source
- Data by AI HUB demo catalogue.
- Size
- 100,000 catalogue target. Sample only is hosted.
- Fields
- id, text, language, domain, split
- Language
- Hausa
- Geography
- Nigeria · Niger
- Limitations
- Tiny authored sample is exposed. The stated 100,000 records are a catalogue target, not files hosted here.
- Potential bias
- Domains skew to markets and greetings. Dialect coverage is not representative of all Hausa varieties.
- Recommended uses
- Interface demos, schema design, teaching annotation.
- Prohibited uses
- Presenting the sample as a production training corpus or reselling it.
- Licence
- research
Versions
v1.0 · 2026-01-15
Initial demo card.
+100000 / −0 · Score set at first review · id, text, language
v1.1 · 2026-03-02
Added domain field and bias note.
+0 / −0 · +1 documentation · Added domain, split
Report dataset
Use this for privacy, copyright, harmful content or a misleading card. Do not attach more personal data than the report needs.
You can also use the report desk.