Hugging Face's logo

Datasets:
michsethowusu
/
twi-sentiments-corpus-v1

Modalities:
Text
Formats:
parquet
Size:
1M - 10M
Libraries:
Datasets
pandas
Dataset card Data Studio Files
xet
Community
1
Dataset Viewer
Auto-converted to Parquet
Twi
stringlengths
1
485
sentiment
stringclasses
2 values
Nyansa mu na woyɛɛ ne nyinaa;
Positive
alo yɛngɛ sone,
Negative
Wosɛe wɔn a wɔnni wo nokorɛ nyinaa.
Negative
Akatua bɛn na ɔde bɛma wɔn a wɔde gyidi som no?
Positive
mepɛ Onyankopɔn ho nimdeɛ sen ɔhyew afɔre.
Positive
na woahu nnebɔneyɛfo asotwe.
Negative
Hwɛ, me nkoa ani begye, na mo de, mo ani bewu.
Negative
Nyansa mu na woyɛɛ ne nyinaa;
Positive
Wɔabɔ dɔnkoro ne fa.
Positive
na woahunu nnebɔneyɛfoɔ asotwe.
Negative
Nhyira ne wɔn a wonhui na wogye di.
Positive
Wosii gyinae sɛ wobebu wɔn ani agu nea wɔn Bɔfo no pɛ so, na wotwaa so aba.
Positive
Mmm hwana ɔbɔɔ realer no,
Negative
Mmm hwana ɔbɔɔ realer no,
Negative
na amumɔyɛfo bɛsan aba wo nkyɛn.
Positive
wo nokwaredi mu, sɛe wɔn.
Negative
ɔbɛdwerɛ ahemfo wɔ nʼabufuhyeɛ da no.
Negative
Nyamenle di nwolɛ ɛzonle ɔ?
Negative
Eyi akyi no, ɔmaa atemmufo dii wɔn so kosii sɛ odiyifo Samuel bae.
Positive
asase so nnipa a wɔn akatua wɔ nkwa yi mu.
Positive
'Teefo na Onyankopɔn ne wɔn di atirimsɛm.'
Positive
Nanso, deɛ ɔsene Salomo no wɔ ha.
Positive
Nanso mo de, nea ɔpɛ sɛ ɔyɛ mo so panyin no nyɛ sɛ mo mu akumaa, na nea odi mo so no nso nyɛ mo mu somfo.
Positive
Saa mfeɛ aduanan yi nyinaa mu, Awurade aka mo ho na hwee ho anhia mo.
Positive
na ɔsram renhyerɛn;
Negative
Mɔlebɛbo ne, ɛnee ɛzoanvolɛma ne ɛnlie ɛnli kɛ Gyisɛse ɔ.
Negative
Dɔnhwere baako pɛ mu, wʼatemmuo aba.'
Negative
na amumɔyɛfo bɛsan aba wo nkyɛn.
Positive
ɔbɛdwerɛw ahemfo wɔ nʼabufuwhyew da no.
Negative
Mede ama Lot asefoɔ sɛ agyapadeɛ."
Positive
Yei nti na da biara wobɛhu Ananse na waka padeɛ mu no.
Positive
Onyankopɔn atirimpɔw a ɔwɔ ma asase ne adesamma no, na minnim ho hwee.
Negative
Emu na ɔtreneeni guan kɔ, na onya ahobammɔ."
Positive
Da biara mu bɔne ankasa dɔɔso ma da no."
Negative
Na mmarima baanu yi de yɛɛ apam.
Positive
Wow akaasoka,
Positive
Monsakra mo adwene na munnye Asɛmpa no nni!"
Positive
na ɛnyɛ akofena anofanu wɔ wɔn nsam,
Positive
Nwoma nko, nyansa nko.
Positive
Yɛ eyinom na wubenya nkwa."
Positive
"O mpaebɔ Tiefo, wo nkyɛn na nnipa a wofi mmaa nyinaa bɛba."
Positive
na mato me bɛmma awowɔ wɔn.
Positive
alo yɛngɛ sone,
Negative
Eyi bɛma mo nnɔbae so ato.
Positive
na wode nneɛma a ɛyɛ duru too yɛn akyi.
Negative
So asase nyinaa temmufo no renyɛ nea ɛteɛ anaa?'
Negative
"Na sɛ mokɔ kurow biara mu na wogye mo fɛw so a, aduan biara a wɔde bɛma mo no, munni.
Positive
Nwoma nko, nyansa nko.
Positive
ne n'asomdwoe no nka yen;
Positive
Da biara mu bɔne ankasa dɔɔso ma da no."
Positive
Kɛ neazo la, Gyihova hanle kɛ, bɛpɛ mrenyia ne mɔ kɔsɔɔti mrenyiazo na ɛnee ɛhye bamaa bɛayɛ bɛtɛɛ wɔ kenle dɔɔnwo anu.
Positive
Eyi bue kwan ma Samariafo pii bɛyɛ gyidifo.
Positive
obiara nni nimdeɛ anaa nhunumu a ɔde bɛka sɛ,
Negative
So wɔyɛ nkurɔfo ayayade wɔ Gehenna?
Negative
Kenkan Onyankopɔn Asɛm da biara da, na wɛn ma mpaebɔ.
Positive
Pam wɔn; esiane wɔn bɔne dodow no nti,
Negative
Nhyira ne wɔn a wonhui na wogye di.
Positive
Na ɛhe ne Yuda sorɔnsorɔmmea?
Positive
Enti wɔnnyɛ agyidifo.
Positive
na amumɔyɛfo bɛsan aba wo nkyɛn.
Positive
Woaka aborɔme bi akyerɛ me nkurɔfo, nanso wonkyerɛɛ me ase."
Negative
Biribiara mu, O Awurade, woyɛ ɔnokwafo.
Positive
Wɔabɔ dɔnkoro ne fa.
Negative
Fa Fa Twins,
Negative
Anadwo yi, wobegye wo kra afi wo nsam, na hena na nneɛma bebree a woapɛ agu hɔ yi, wode begyaw no?'
Negative
Wɔabɔ dɔnkoro ne fa.
Positive
"Yi da bi to hɔ a wɔde bɛyɛ akɔnkyen, na ma Nabot ntena anuonyambea wɔ nnipa no mu.
Positive
Nhyira ne wɔn a wonhui na wogye di.
Positive
nʼanwanwadeɛ akyi no, wɔannye anni.
Negative
Na ma wɔn ko-ma nni a-h'ru-si.
Negative
Obi wɔ mo mu a onim nyansa na ɔwɔ ntease?
Positive
ogya a ɛhyew nneɛma di nʼanim,
Positive
ogya a ɛhyew nneɛma di nʼanim,
Positive
mpo ɔbɛyɛ yɛn kwankyerɛfoɔ akɔsi awieeɛ.
Positive
"Mɛbɔ wɔn ho ban afi wɔn a wɔhaw wɔn no ho."
Positive
Wo a wutwa nkontompo nko ara.
Negative
Onyankopɔn Asɛm ka sɛ: "Noa ne nokware Nyankopɔn nantewee."
Positive
Nyansa mu na woyɛɛ ne nyinaa; w'abɔde ahyɛ asase so ma.
Positive
Sɛdeɛ Atwerɛsɛm no ka no, ɔteneneeni firi gyidie mu bɛnya nkwa.
Positive
Odiyifo bi wɔ hɔ a mo agyanom antaa no anaa?
Negative
ne Yuda nkurow nyinaa so.
Positive
Na nkɔmhyɛni no ne n'apamfoɔ pii ahyɛ afiase abosome bebree.
Negative
Nhyira ne wɔn a wonhui na wogye di.
Positive
Nhyira nka wɔn a wotie adiyisɛm a ɛwɔ saa nhoma yi mu no!"
Positive
Na afei ɛdeɛn na wobɛyɛ de ahyɛ wo din kɛseɛ a ɛwɔ animuonyam no?"
Positive
Onyankopɔn biara nni hɔ te sɛ ɔno.'
Negative
"Saa nsɛnkyerɛnne yi bedi wɔn a wogye me di no akyi.
Positive
Anaa Onyankopɔn na ɔbɔɔ no?
Negative
Anaa Onyankopɔn na ɔbɔɔ no?
Negative
Enti ɔkyerɛwee sɛ: "Wo a wokyerɛkyerɛ obi no, wonkyerɛkyerɛ wo ho?"
Negative
"Ne honhom fi ne mu, na ɔsan kɔ dɔte mu; da no ara, ne nsusuwii yera."
Negative
Adɛn nti na Onyankopɔn bɔɔ no?"
Negative
Sɛ meka asase yi so nsɛm kyerɛ mo na munnye nni a, ɛbɛyɛ dɛn na sɛ meka ɔsoro nsɛm kyerɛ mo a, mubegye adi?
Negative
'Teefo na Onyankopɔn ne wɔn di atirimsɛm.'
Positive
wo nsam wɔ ahoɔden, wama wo nsa nifa so.
Positive
deɛ ɔnwonoo wo, na ɔyɛɛ wo wɔ awotwaa mu,
Positive
Na sɛ wutie m'asɛm a, wubenyin akyɛ paa.'
Negative
na mato me bɛmma awowɔ wɔn.
Positive
Nanso Onyankopɔn bɔɔ saa nnuan no sɛ, sɛ gyidifo a wɔahu nokware no nsa ka na wɔbɔ so mpae a, wotumi di.
Negative
Wosɛe wɔn a wɔnni wo nokorɛ nyinaa.
Negative
End of preview. Expand in Data Studio

Twi Sentiments Dataset

This dataset contains over 1.3 million Twi sentences with binary sentiment classifications, designed to support the development of robust sentiment analysis models for the Twi language.

Dataset Information

  • Language: Twi (Akan)
  • Task: Binary sentiment classification
  • Size: 1,383,615 examples
  • Format: CSV
  • Columns:
    • Twi: Text in Twi language
    • sentiment: Binary sentiment label (positive/negative)

Dataset Statistics

  • Total examples: 1,383,615
  • Dataset size: 76.8 MB
  • Download size: 46.0 MB
  • Split: Training set only

Usage Example 

from datasets import load_dataset

# Load the dataset from the Hugging Face Hub
dataset = load_dataset("michsethowusu/twi-sentiments-corpus")

# Access the training split
train_data = dataset["train"]

# Print first example
print(f"Text: {train_data[0]['Twi']}")
print(f"Sentiment: {train_data[0]['sentiment']}")

# Get dataset info
print(f"Total examples: {len(train_data)}")

Applications

  • Sentiment Analysis: Train models to classify sentiment in Twi text
  • African Language NLP: Support research in low-resource language processing
  • Cross-lingual Studies: Compare sentiment patterns across languages
  • Social Media Analysis: Analyze sentiment in Ghanaian social media content
  • Opinion Mining: Extract opinions from Twi text data

Data Quality

This corpus has been curated to provide a substantial resource for Twi sentiment analysis research. The dataset aims to capture diverse expressions of sentiment in the Twi language to enable the creation of robust and generalizable models.

Citation

If you use this dataset in your research, please cite:

@dataset{twi_sentiments_2024,
  title={Twi Sentiments Corpus: A Large-Scale Sentiment Classification Dataset for Twi},
  author={[Author Name]},
  year={2024},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/michsethowusu/twi-sentiments-corpus}
}

License

This dataset is made available for research purposes. Please ensure proper attribution when using this dataset in your work.

Contact

For questions or issues regarding this dataset, please open an issue on the dataset repository or contact the maintainers.

Contributing to African Language NLP Research

This dataset is part of ongoing efforts to advance natural language processing capabilities for African languages, particularly Twi. I encourage researchers to use this resource to develop better sentiment analysis tools for the Twi-speaking community.

Downloads last month
21
Size of downloaded dataset files:
46 MB
Size of the auto-converted Parquet files:
46 MB
Number of rows:
1,383,615

Collection including michsethowusu/twi-sentiments-corpus-v1