Dataset Card for AfriSenti Dataset

Dataset Summary

AfriSenti is the largest sentiment analysis dataset for under-represented African languages, covering 110,000+ annotated tweets in 14 African languages (Amharic, Algerian Arabic, Hausa, Igbo, Kinyarwanda, Moroccan Arabic, Mozambican Portuguese, Nigerian Pidgin, Oromo, Swahili, Tigrinya, Twi, Xitsonga, and Yoruba).

The datasets are used in the first Afrocentric SemEval shared task, SemEval 2023 Task 12: Sentiment analysis for African languages (AfriSenti-SemEval). AfriSenti allows the research community to build sentiment analysis systems for various African languages and enables the study of sentiment and contemporary language use in African languages.

Supported Tasks and Leaderboards

The AfriSenti can be used for a wide range of sentiment analysis tasks in African languages, such as sentiment classification, sentiment intensity analysis, and emotion detection. This dataset is suitable for training and evaluating machine learning models for various NLP tasks related to sentiment analysis in African languages. SemEval 2023 Task 12 : Sentiment Analysis for African Languages

Languages

14 African languages (Amharic (amh), Algerian Arabic (ary), Hausa(hau), Igbo(ibo), Kinyarwanda(kin), Moroccan Arabic/Darija(arq), Mozambican Portuguese(por), Nigerian Pidgin (pcm), Oromo (oro), Swahili(swa), Tigrinya(tir), Twi(twi), Xitsonga(tso), and Yoruba(yor)).

Dataset Structure

Data Instances

For each instance, there is a string for the tweet and a string for the label. See the AfriSenti dataset viewer to explore more examples.

{
  "tweet": "string",
  "label": "string"
}

Data Fields

The data fields are:

tweet: a string feature.
label: a classification label, with possible values including positive, negative and neutral.

Data Splits

The AfriSenti dataset has 3 splits: train, validation, and test. Below are the statistics for Version 1.0.0 of the dataset.

ama	arq	hau	ibo	ary	orm	pcm	pt-MZ	kin	swa	tir	tso	twi	yo
train	5,982	1,652	14,173	10,193	5,584	-	5,122	3,064	3,303	1,811	-	805	3,482	8,523
dev	1,498	415	2,678	1,842	1,216	397	1,282	768	828	454	399	204	389	2,091
test	2,000	959	5,304	3,683	2,962	2,097	4,155	3,663	1,027	749	2,001	255	950	4,516
total	9,483	3,062	22,155	15,718	9,762	2,494	10,559	7,495	5,158	3,014	2,400	1,264	4,821	15,130

How to use it

from  datasets  import  load_dataset

# you can load specific languages (e.g., Amharic). This download train, validation and test sets. 
ds = load_dataset("shmuhammad/AfriSenti-twitter-sentiment", "amh")

# train set only
ds = load_dataset("shmuhammad/AfriSenti-twitter-sentiment", "amh", split = "train")

# test set only
ds = load_dataset("shmuhammad/AfriSenti-twitter-sentiment", "amh", split = "test")

# validation set only
ds = load_dataset("shmuhammad/AfriSenti-twitter-sentiment", "amh", split = "validation")

Dataset Creation

Curation Rationale

AfriSenti Version 1.0.0 aimed to be used in the first Afrocentric SemEval shared task SemEval 2023 Task 12: Sentiment analysis for African languages (AfriSenti-SemEval) .

Source Data

Twitter

Initial Data Collection and Normalization

[More Information Needed]

Who are the source language producers?

[More Information Needed]

Annotations

Annotation process

[More Information Needed]

Who are the annotators?

[More Information Needed]

Personal and Sensitive Information

We anonymized the tweets by replacing all @mentions by @user and removed all URLs.

Considerations for Using the Data

Social Impact of Dataset

The Afrisenti dataset has the potential to improve sentiment analysis for African languages, which is essential for understanding and analyzing the diverse perspectives of people in the African continent. This dataset can enable researchers and developers to create sentiment analysis models that are specific to African languages, which can be used to gain insights into the social, cultural, and political views of people in African countries. Furthermore, this dataset can help address the issue of underrepresentation of African languages in natural language processing, paving the way for more equitable and inclusive AI technologies.

[More Information Needed]

Discussion of Biases

[More Information Needed]

Other Known Limitations

[More Information Needed]

Additional Information

Dataset Curators

AfriSenti is an extension of NaijaSenti, a dataset consisting of four Nigerian languages: Hausa, Yoruba, Igbo, and Nigerian-Pidgin. This dataset has been expanded to include other 10 African languages, and was curated with the help of the following:

Language	Dataset Curators
Algerian Arabic (arq)	Nedjma Ousidhoum, Meriem Beloucif
Amharic (ama)	Abinew Ali Ayele, Seid Muhie Yimam
Hausa (hau)	Shamsuddeen Hassan Muhammad, Idris Abdulmumin, Ibrahim Said, Bello Shehu Bello
Igbo (ibo)	Shamsuddeen Hassan Muhammad, Idris Abdulmumin, Ibrahim Said, Bello Shehu Bello
Kinyarwanda (kin)	Samuel Rutunda
Moroccan Arabic/Darija (ary)	Oumaima Hourrane
Mozambique Portuguese (pt-MZ)	Felermino Dário Mário António Ali
Nigerian Pidgin (pcm)	Shamsuddeen Hassan Muhammad, Idris Abdulmumin, Ibrahim Said, Bello Shehu Bello
Oromo (orm)	Abinew Ali Ayele, Seid Muhie Yimam, Hagos Tesfahun Gebremichael, Sisay Adugna Chala, Hailu Beshada Balcha, Wendimu Baye Messell, Tadesse Belay
Swahili (swa)	Davis Davis
Tigrinya (tir)	Abinew Ali Ayele, Seid Muhie Yimam, Hagos Tesfahun Gebremichael, Sisay Adugna Chala, Hailu Beshada Balcha, Wendimu Baye Messell, Tadesse Belay
Twi (twi)	Salomey Osei, Bernard Opoku, Steven Arthur
Xithonga (tso)	Felermino Dário Mário António Ali
Yoruba (yor)	Shamsuddeen Hassan Muhammad, Idris Abdulmumin, Ibrahim Said, Bello Shehu Bello

Licensing Information

This AfriSenti is licensed under a Creative Commons Attribution 4.0 International License

Citation Information

@inproceedings{Muhammad2023AfriSentiAT,
  title={AfriSenti: A Twitter Sentiment Analysis Benchmark for African Languages},
  author={Shamsuddeen Hassan Muhammad and Idris Abdulmumin and Abinew Ali Ayele and Nedjma Ousidhoum and David Ifeoluwa Adelani and Seid Muhie Yimam and Ibrahim Sa'id Ahmad and Meriem Beloucif and Saif Mohammad and Sebastian Ruder and Oumaima Hourrane and Pavel Brazdil and Felermino D'ario M'ario Ant'onio Ali and Davis Davis and Salomey Osei and Bello Shehu Bello and Falalu Ibrahim and Tajuddeen Gwadabe and Samuel Rutunda and Tadesse Belay and Wendimu Baye Messelle and Hailu Beshada Balcha and Sisay Adugna Chala and Hagos Tesfahun Gebremichael and Bernard Opoku and Steven Arthur},
  year={2023}
}

@article{muhammad2023semeval,
  title={SemEval-2023 Task 12: Sentiment Analysis for African Languages (AfriSenti-SemEval)},
  author={Muhammad, Shamsuddeen Hassan and Abdulmumin, Idris and Yimam, Seid Muhie and Adelani, David Ifeoluwa and Ahmad, Ibrahim Sa'id and Ousidhoum, Nedjma and Ayele, Abinew and Mohammad, Saif M and Beloucif, Meriem},
  journal={arXiv preprint arXiv:2304.06845},
  year={2023}
}

Contributions

[More Information Needed]