Datasetpublic

NjeriKahoro/anv-kikuyu-banking-subset

Anv-Kikuyu Banking Subset A domain-filtered subset of Kikuyu (Gĩkũyũ) speech data focused on banking and financial-transaction content, combined into a single repository with train, test, and validation splits. Source This dataset is a filtered subset of Anv-ke/kikuyu, part of the African Next Voices (ANV) collection. All audio, transcriptions, and underlying speaker data originate from that source dataset. Full credit for data collection belongs to the… See the full description on the dataset page: https://huggingface.co/datasets/NjeriKahoro/anv-kikuyu-banking-subset.

sourceHugging Facecc-by-4.0updated 2mo agoView on Hugging Face

1likes93downloads

Dataset Card

Anv-Kikuyu Banking Subset

A domain-filtered subset of Kikuyu (Gĩkũyũ) speech data focused on banking and financial-transaction content, combined into a single repository with train, test, and validation splits.

Source

This dataset is a filtered subset of [Anv-ke/kikuyu](https://huggingface.co/Anv-ke/kikuyu), part of the African Next Voices (ANV) collection. All audio, transcriptions, and underlying speaker data originate from that source dataset. Full credit for data collection belongs to the Anv-ke/African Next Voices project.

Rows here were selected by scanning the source dataset for banking/financial-domain content using a keyword filter (e.g. terms relating to banks (bengi), ATMs, and accounts(akaunti, akauti) in Kikuyu and English) applied to the transcription field.

Dataset Structure

The three splits mirror the source dataset's own train, test, and validation splits — rows were filtered independently from each corresponding split in Anv-ke/kikuyu, not pooled and randomly re-divided. | Split | Rows | |-------|------| | train | 820 | | test | 76 | | validation | 164 |

Columns

Column	Description
`audio`	The speech audio clip
`filename`	Original filename from the source dataset
`type`	Recording type (e.g. scripted/unscripted)
`split`	Original split label from the source dataset
`recorder_uuid`	Anonymized recorder/speaker identifier, as provided by the source
`domain`	Content domain label, as provided by the source
`transcription`	Kikuyu transcription of the audio
`language`	Language code
`source_repo`	Which intermediate filtering repo this row was pulled from during construction of this combined dataset

Processing Notes

—Domain filtering: Rows were kept if their transcription matched a keyword filter targeting banking/financial content.

Intended Use

Intended for research and development of ASR and related speech technology for Kikuyu, particularly in the banking/financial domain.

License and Attribution

Please refer to and comply with the license and usage terms of the original Anv-ke/kikuyu dataset, which governs the underlying audio and speaker data. If you use this dataset, please cite the original Anv-ke/African Next Voices dataset as the primary data source, alongside this repository for the domain-filtering methodology.