News Article Dataset Labeled for Political Bias

CS 230 News V3 Test Liberal contains text data for classifying news articles, likely as part of a Stanford University CS230 course project. The dataset was created by myradeng and last updated on Hugging Face in May 2022. Platform tags indicate it contains between 10,000 and 100,000 rows of tabular and text data from the US region.

Use Cases

• Train a classifier to predict a 'bias' or 'stance' label from article text content.
• Analyze linguistic features (e.g., word choice, sentiment) in text to correlate with political leaning labels.
• Benchmark transformer models like BERT on a text classification task using the provided labels and article text.
• Perform exploratory data analysis on the distribution of labels across the dataset's news sources or time periods.

Strengths

• Dataset size is categorized between 10,000 and 100,000 rows, providing a substantial sample.
• Includes both tabular metadata and raw text modality for flexible analysis.

Limitations

• Data is over two years old, limiting relevance to current media trends.
• The specific definition of the 'liberal' test label and potential class imbalance are unknown.

Provenance

Source

Likely a Stanford CS230 course project, hosted by author myradeng on Hugging Face.

Collection Method

Presumably collected and labeled from online news sources; method unknown.

Time Range

null

Freshness

Last updated in May 2022; no update frequency specified.

Geography

Platform tags suggest a primary focus on the United States (Regionus).