News Bias Detection Dataset for Media Analysis

A text dataset for detecting bias in news articles, curated by author cmpatino and hosted on Hugging Face. The dataset was last updated in December 2025. It contains news articles optimized for analysis with libraries like Polars and Pandas.

Use Cases

• Train a classifier to predict bias labels from article text content.
• Analyze linguistic patterns and sentiment in news articles to quantify slant.
• Perform comparative media analysis using the 'Regionus' tag to study U.S. news sources.
• Fine-tune language models for specific bias detection tasks using the provided text modality.

Strengths

• Dataset is optimized for processing, indicated by the 'OPTIMIZED-PARQUET' tag.
• Supports multiple data analysis libraries (Polars, Pandas, Datasets).
• Last update was recent (December 2025).

Limitations

• Unknown sample size and number of articles limits reproducibility assessment.
• Lack of column details prevents understanding of specific bias labels or features.
• Geographic coverage may be biased towards U.S. ('Regionus') sources.

Provenance

Source

Hugging Face, author cmpatino.

Collection Method

Collection method for news articles is unknown.

Time Range

null

Freshness

Updated December 2025.

Geography

Primarily United States, based on platform tag 'Regionus'.