2.7 Million News Articles from 27 US Publications

Aggregating 2.7 million news articles and essays. The articles are primarily from 27 American publications and mostly span from 2016 to early 2020.

Use Cases

• Analyze article text and publication name to study media bias across 27 US publications.
• Perform temporal analysis on article content using the date, year, and month fields to track topic evolution from 2016 to 2020.
• Conduct publication-level comparisons by aggregating articles based on the publication name field.
• Train a text classification model using article text and associated metadata like date and publication.

Strengths

• Contains 2.7 million individual articles and essays.
• Articles are sourced from 27 distinct American publications.
• Provides temporal coverage from 2016 to early 2020.

Limitations

• The specific column structure and available features are not detailed.
• Data is not recent, with the last articles from early 2020.
• Geographic focus is limited to the United States.

Provenance

Source

Andrew Thompson / rjac on Hugging Face.

Collection Method

null

Time Range

Mostly 2016 to early 2020.

Freshness

null

Geography

United States.