The Ghana-QA Conversations dataset consists of question-answer pairs from Ghanaian news articles formatted as conversational exchanges. It contains questions in three widely spoken Ghanaian languages with their corresponding answers in a chat-style JSON format.

Total Conversations: 3,547,660
Languages: Twi, Ewe, Ga
Format: JSON conversations with human/assistant turns
Source: Ghanaian news articles
Project: Part of the GhanaNLP initiative

Data Format

Each row contains a conversations field with JSON-formatted dialogue:

[
  {"from": "human", "value": "Question text here"},
  {"from": "gpt", "value": "Answer text here"}
]

How to Use

Loading with Hugging Face Datasets

from datasets import load_dataset
import json

# Load the dataset
dataset = load_dataset("ghananlpcommunity/ghana-qa-conversations")

# Parse a conversation
conversation = json.loads(dataset['train'][0]['conversations'])
print(f"Human: {conversation[0]['value']}")
print(f"Assistant: {conversation[1]['value']}")

Using with Pandas

import pandas as pd
import json

# Load the CSV
df = pd.read_csv("data.csv")

# Parse conversations
df['parsed_conversations'] = df['conversations'].apply(json.loads)

# Extract questions and answers
df['question'] = df['parsed_conversations'].apply(lambda x: x[0]['value'])
df['answer'] = df['parsed_conversations'].apply(lambda x: x[1]['value'])

print(df[['question', 'answer']].head())

Fine-tuning Chat Models

This format is ideal for fine-tuning conversational AI models:

from datasets import load_dataset

dataset = load_dataset("ghananlpcommunity/ghana-qa-conversations")

# The conversations are ready to use with chat model training pipelines
# Compatible with formats used by ChatGPT, Llama, Mistral, etc.

Use Cases

Chat model fine-tuning: Train conversational AI on Ghanaian languages
Question-answering systems: Build QA systems for low-resource languages
Multilingual dialogue research: Study conversational patterns across languages
Cross-lingual transfer learning: Improve model performance on African languages

Language Distribution

Twi (tw): Akan language spoken in Ghana
Ewe (ee): Niger-Congo language spoken in Ghana and Togo
Ga (gaa): Kwa language spoken primarily in the Greater Accra Region

Citation

If you use this dataset, please cite:

@dataset{Ghana-QA-Conversations_2025,
  title = {Ghana-QA Conversations: Ghanaian Question-Answer Pairs in Chat Format},
  author = {Mich-Seth Owusu and Priscilla Lartey and Gerhardt Datsomor},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/ghananlpcommunity/ghana-qa-conversations}
}

License

This dataset is licensed under CC-BY-NC-4.0 (Creative Commons Attribution-NonCommercial 4.0 International).

Acknowledgments

Special thanks to the primary contributors: Mich-Seth Owusu, Priscilla Lartey, Gerhardt Datsomor, and all volunteers from the GhanaNLP community. This work is part of the GhanaNLP initiative to advance natural language processing for Ghanaian languages.

Related Datasets

ghananlpcommunity/ghana-qa - Original format with separate question/answer columns

Contact

For questions or feedback, please reach out through the GhanaNLP community channels or open an issue in the dataset repository.

Downloads last month: 16