Datasets
DATASETHugging Face
5/5

WaxalNLP - large-scale African speech corpus (google/WaxalNLP), dga_asr subset (191,000 utterances)

Southern Dagaare(dga)SpeechGender dataWest Africa
View Source

Dataset Info

Language
Southern Dagaare (dga)
Modality
Speech
Domain
General
Size
191000
Institution
Google Research + Gates Foundation; Dagaare data by University of Ghana (UGSpeechData); arXiv:2602.02734
Last Updated
2 Jul 2026

License & Access

SPDX identifierCC-BY-4.0

Open — attribution required; commercial use allowed.

RedistributionYes

PII RISK

Low risk

Full Record

34 of 46 fields stated

Every column the registry tracks. Values marked derivedwere filled by a rule or read from the source's text rather than stated in its metadata; the rule or the quoted sentence is shown under the value.

Registry columns (25/27)

iso code
dga
Dataset name
WaxalNLP - large-scale African speech corpus (google/WaxalNLP), dga_asr subset (191,000 utterances)
Modality
Speech
Type
dataset
Link
https://huggingface.co/datasets/google/WaxalNLP
NLP Task
ASR
Domain
General
Data Format
parquet
Size Metric
Files
Size Value
191000
License
CC-BY-4.0
Open Access
Open
Source Repository
Hugging Face
Year Published
2026
Creator/Author(s)
Google Research + Gates Foundation; Dagaare data by University of Ghana (UGSpeechData); arXiv:2602.02734
Funder/Partner
Google
Dialect Coverage
Standard
Gender Balance
Yes
Speaker Count
Not stated
Annotation Type
Human
Status
Active
Has Datasheet
Yes
PII Risk
Low
Consent Verified
Not stated
Verified By
Agboola Blessing
Date Indexed
2026-07-02
Notes / Flags
CONFIRMED headline resource - the largest speech corpus for any language in this project. WaxalNLP lists subset 'dga_asr' with 191,000 rows (verified on the dat…

Quality (DQI) (6/12)

Datasheet Fields Completed
Not stated
SPDX License ID
CC-BY-4.0
Redistribution Allowed?
Yes
Content Hash
5f4d8ca24f2b9d168b2ee545f1febaaff4b40580derivedHugging Face card: commit sha
Submitted By
hf_data pipeline (automated)
Fetch Date
2026-07-02
Last Updated Date
2026-09-01derivedHugging Face card: lastModified
Annotation Guidelines Available?
Not stated
IAA Score
Not stated
Cleaning Notes
Not stated
Benchmark Exists?
Not stated
Benchmark Citation
Not stated

Representativeness (DRI) (3/7)

Gender Distribution
Not stated
Region Coverage
Ghana, Senegal
Region Distribution
Not stated
Age Range Coverage
Not stated
Register
Spoken
Marginalized Group Note
Not stated
Demographic Fields Reported Count
1

Readiness Score Breakdown

5/5
Data Coverage100%
Model Availability100%
Benchmarks100%
Documentation71%
Recency100%

Representativeness

Gender Balance
Regional Distribution
Dialect Coverage
Inclusivity

Provenance

Source URL
https://huggingface.co/datasets/google/WaxalNLP
Fetch Timestamp
2 Jul 2026
Content Hash
5f4d8ca24f2b9d168b2ee545f1febaaff4b40580

Audit Log

Record indexed from Hugging Face

2 Jul 2026

Source status: Active

2 Jul 2026

Other Copies

The same resource, also published at these locations. It is counted once.

dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

Data format: parquet

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026
dataset

WaxalNLP

automatic-speech-recognition, text-to-speech · speech domain · isaacoluwafemiog

8 languages15 copiesSouthern Dagaare(dga)Audioautomatic-speech-recognition, text-to-speech
HUHugging Face
Open27 Oct 2026

Related Datasets

dataset

Dagaare Wikipedia (dga.wikipedia.org) - full wiki dump (wikimedia/wikipedia & mirrors)

CONFIRMED (edition exists; verify snapshot). The Dagaare Wikipedia went live at dga.wikipedia.org in Dec 2023 (the 3rd Mabia language on Wikipedia). Its arti…

49 languages32 copiesSouthern Dagaare(dga)TextLM
HUHugging Face
Open2 Jul 2026
dataset

FineWeb-2 - multilingual web corpus (HuggingFaceFW/fineweb-2), dga_Latn subset

VERY LIKELY (verify). FineWeb-2 covers 1,868 language-script pairs via GlotLID. Southern Dagaare (~700k-1.3M speakers, language of instruction in Ghana, own …

517 languages2 copiesSouthern Dagaare(dga)TextLM
HUHugging Face
Open2 Jul 2026
dataset

DCAD-2000 - Multilingual Common-Crawl corpus (openbmb/DCAD-2000), dga_Latn subset

VERY LIKELY (verify). DCAD-2000 (2000+ langs from FineWeb/Fineweb-2 + MaLA + Common Crawl). If dga is in FineWeb-2 (expected), a 'dga_Latn' subset follows. F…

577 languagesSouthern Dagaare(dga)TextLM
HUHugging Face
Open2 Jul 2026