File size: 4,658 Bytes
2238639
ad361a7
 
2238639
ad361a7
 
 
fbc7587
 
 
 
 
 
 
 
 
 
 
 
ad361a7
 
 
 
 
 
fbc7587
 
ad361a7
8678c26
ad361a7
 
fbc7587
 
 
 
 
71a7f6d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e8c6432
71a7f6d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e8c6432
71a7f6d
 
 
e8c6432
71a7f6d
 
e8c6432
71a7f6d
 
 
 
 
e8c6432
71a7f6d
 
 
e8c6432
71a7f6d
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
---
language:
- ve
license: cc-by-nc-sa-4.0
size_categories:
- 1K<n<10K
pretty_name: ZaBantu News Headlines[RAW]
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: text
    dtype: string
  - name: language
    dtype: string
  - name: word_count
    dtype: int64
  - name: char_count
    dtype: int64
  - name: sentence_count
    dtype: int64
  - name: PotentialSplits
    sequence: string
  - name: md5
    dtype: string
  splits:
  - name: train
    num_bytes: 8799449
    num_examples: 16607
  download_size: 5317813
  dataset_size: 8799449
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---


# ZaBantu News[Raw]

## Table of Contents
- [Dataset Description](#dataset-description)
    - [Dataset Summary](#dataset-summary)
    - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
    - [Languages](#languages)
- [Dataset Structure](#dataset-structure)
    - [Data Instances](#data-instances)
    - [Data Fields](#data-fields)
    - [Data Splits](#data-splits)
- [Dataset Creation](#dataset-creation)
    - [Curation Rationale](#curation-rationale)
    - [Source Data](#source-data)
    - [Annotations](#annotations)
- [Considerations for Using the Data](#considerations-for-using-the-data)
    - [Social Impact of Dataset](#social-impact-of-dataset)
    - [Licensing Information](#licensing-information)
    - [Citation Information](#citation-information)

## 1. Dataset Description

### Dataset Summary
ZaBantu News[Raw] is a collection of news headlines in Sepedi (Northern Sotho) and Tshivenda languages, sourced from various public media platforms. This dataset is designed to support NLP research and development in low-resource languages, providing raw, unlabelled headlines. It complements the "ZaBantu News" dataset, which includes a mix of machine and human labels following [IPTC news codes](https://www.iptc.org/std/NewsCodes/treeview/mediatopic/mediatopic-en-GB.html).

### Supported Tasks and Leaderboards
This dataset is suitable for tasks such as language modeling, unsupervised learning, and pre-training for downstream NLP tasks (e.g., text classification, topic modeling). It can also be used for linguistic studies and developing NLP tools for Sepedi and Tshivenda.

### Languages
The dataset contains text in Sepedi (Northern Sotho) and Tshivenda, both of which are low-resource languages spoken in South Africa.

## 2. Dataset Structure

### Data Instances
A data instance in "ZaBantu News[Raw]" consists of a single news headline. Some headlines may be segmented using delimiters like asterisks or full stops for further breakdown.

### Data Fields
- `headline`: the news headline text.
- `language`: language of the headline (Sepedi or Tshivenda).

### Data Splits
The dataset is currently not split into standard subsets (e.g., training, validation, testing). Users may consider splitting the dataset according to their needs for specific tasks.

## 3. Dataset Creation

### Curation Rationale
The dataset was curated to address the scarcity of digital resources in Sepedi and Tshivenda for NLP research. It aims to facilitate the development of language models and other NLP tools for these languages.

### Source Data
#### Initial Data Collection and Normalization
Headlines were collected from various public media platforms, ensuring a diverse representation of topics and styles. 

#### Who are the source language producers?
The source language producers include journalists and content creators from multiple media outlets that publish news in Sepedi and Tshivenda.

### Annotations
#### Annotation process
The dataset is currently not annotated. It serves as a raw collection of news headlines for potential use in unsupervised learning tasks or as a foundation for future annotated datasets.

#### Who are the annotators?
Not applicable.

## 4. Motivation

- **Social Impact of Dataset** - The dataset contributes to NLP research in low-resource languages, potentially aiding in the development of technology that supports linguistic diversity and inclusion.

## 5. Licensing Information
The dataset is shared under a Share Alike license, allowing for adaptation and distribution of derivative works under the same or a compatible license.

## 6. Citation Information
Please cite this dataset as follows:

```bibtext
@misc{zabantu_news_raw,
author = {Ndamulelo Nemakhavhani},
title = {ZaBantu News[Raw]: Collecting Tshivenda and Sepedi News Headlines for Cross-Lingual Transfer Evaluation,
year = {2024},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://huggingface.co/datasets/ndamulelonemakh/zabantu-news}}
}
```