708,241 English language news articles collected from global news sites between January 2017 and December 2019. The collection is derived from the Common Crawl CC-NEWS bucket and processed using the news-please extraction tool to provide structured text and metadata.
Use Cases
- Train language models on the article text content
- Analyze media trends using the publication date metadata from 2017 to 2019
- Validate news extraction pipelines by comparing outputs to the news-please processed data
- Perform sentiment analysis on global news events occurring between 2017 and 2019
Strengths
- 708,241 English language news articles
- Temporal coverage spanning from January 2017 to December 2019
- Data extracted from the Common Crawl CC-NEWS bucket on AWS S3
- Processed using the news-please integrated web crawler and information extractor