metadata

license: apache-2.0
language:
  - en
pretty_name: TheGuardian.com Website Articles
tags:
  - webdataset
  - News
  - Articles
  - Text Classification
  - Information Extraction
  - Natural Language Processing
  - Web Scraping
  - Data Curation
  - English
  - Multilingual
  - Text Analysis
  - Sentiment Analysis
  - Topic Modeling
  - Machine Learning
  - Data Science
  - Apache-2.0
  - Research
  - Journalism
  - Media
  - Large-Scale Data
task_categories:
  - text-classification
  - summarization
  - feature-extraction
size_categories:
  - 100M<n<1B

Dataset Creation

Curation Rationale

The dataset was curated to facilitate research and development in natural language processing tasks such as text classification and information extraction from news articles.

Source Data

Initial Data Collection and Normalization

Articles and similar content were scraped from theguardian.com website.

Encoding

The primary language of the dataset is English, but it may contain content in other languages. It is recommended to handle the data in UTF-8 encoding to accommodate the full range of characters.

Dataset Structure

Data Instances

A typical data row contains:

URL: [string] The URL of the scraped article.
Article Category: [string] The category assigned to the article.
Publication Date: [datetime] The date and time the article was published.
Article Author: [string] The author of the article.
Article Title: [string] The title of the article.
Article Contents: [string] The full text content of the article.
Data Quality: [string] 'Full' if all data from the article was scraped, 'Partial' if some data may be missing.

Dataset Statistics

Tokens

The dataset model includes an estimated 135,808,383 tokens.

Data Quality Distribution

Number of rows with 'Full' data quality: 87,641
Number of rows with 'Partial' data quality: 50,594

Who are the intended users?

Researchers, data scientists, and developers interested in text analysis of news articles.

Considerations

Content Warning

The dataset may contain explicit content or material that is not considered child-friendly. Users should exercise caution and review the content before using it in applications accessible by a younger audience.

Potential Biases

The dataset reflects the content as published on theguardian.com and may inherently contain biases present in the source material. Users should be aware of these potential biases when analyzing the dataset and consider them in their research or applications.

Legal and Ethical Considerations

Dataset Licensing

The dataset is distributed under the Apache License 2.0, which allows for a wide range of uses.

Data Privacy

The dataset contains publicly available information from articles. However, users should ensure they comply with The Guardian's terms of use and privacy policies when using this data.

Content Usage Restrictions

According to the robots.txt file of theguardian.com, the content is made available under their terms and conditions of use. The following uses are not permitted:

For large language models (LLMs)
For machine learning and/or artificial intelligence-related purposes
With any of the aforementioned technologies
For any commercial purposes

For further assistance or permissions, contact licensing@theguardian.com.

Who are the dataset creators?

The dataset model was created by Stefan Carter (stefan171@gmail.com).