French COVID-19 News Articles From 50+ Outlets

Comprising just over 40,000 unique French-language news articles from more than 50 different French-speaking online newspapers. It was prepared using the news-please web crawler and information extractor. The dataset supports abstractive summarization and topic classification tasks.

Use Cases

• Train a topic classification model on French news articles to categorize COVID-19 related content.
• Perform abstractive summarization on the French news articles to generate concise article summaries.
• Analyze the distribution of news sources across the more than 50 French-speaking online newspapers.
• Fine-tune a French language model for text generation tasks using the 40k news articles.

Strengths

• Just over 40,000 unique news articles provide a substantial text corpus.
• Articles are sourced from more than 50 different French-speaking online newspapers, offering source diversity.
• Data was prepared using news-please, an integrated web crawler and information extractor designed for news.

Limitations

• The dataset card is noted as unfinished, indicating potential incomplete documentation.
• Specific column structure, sample data, and file formats are unknown, complicating initial analysis.
• The temporal and geographic coverage of the articles is not specified.

Provenance

Source

More than 50 different French-speaking online newspapers.

Collection Method

Prepared using news-please, an integrated web crawler and information extractor for news.

Freshness

The dataset was last updated on 2024-02-23.

Geography

French-speaking sources, specific coverage unknown.