Turkish Media Archive Dataset

Columns

• date: This column specifies the date each article was published, presented in Turkish text format.
• author: Contains the name of the individual author responsible for writing the article.
• title: Represents the heading or title of the news article.
• link: Provides the direct web address (URL) to the original article as published on the Sabah newspaper's website.
• text: Encompasses the full body content of the news article.

Distribution

• Format: The dataset is provided as a CSV file.
• Size: It comprises 2415 individual articles (rows) and 5 distinct columns.
• Structure:
  • There are 204 unique authors contributing to the dataset.
  • The collection features 2405 unique article titles.
  • Each article has a unique link, totalling 2415 unique links.
  • The dataset contains 2413 unique article texts, indicating a high degree of distinct content.
  • A separate stop-word.txt file, consisting of 1797 words, is included for use with the Zemberek Turkish natural language processing library.

Usage

• Text Classification: Training models to categorise Turkish news articles.
• Natural Language Processing (NLP): Developing, testing, and refining NLP models tailored for the Turkish language.
• Sentiment Analysis: Analysing the emotional tone and public opinion within Turkish news content.
• Author Attribution: Investigating and identifying authors based on their unique writing styles.
• Trend Analysis: Monitoring evolving themes and topics covered in Turkish media over time.
• Topic Modelling: Discovering and extracting key thematic structures from a large corpus of news articles.

Coverage

• Geographic: The articles originate from the Sabah newspaper, making the primary geographic focus Turkey.
• Time Range: The dataset covers a five-year period, with articles published between 2017 and 2021.
• Demographic: The content is sourced from articles written by 204 different authors.

License

Who Can Use It

• Academic Researchers: Conducting studies in linguistics, media studies, political science, or social sciences related to Turkish content.
• Data Scientists: Building and deploying machine learning models for text processing and analysis.
• NLP Engineers and Developers: Enhancing or creating tools and applications for Turkish language understanding.
• Media Analysts: Gaining insights into news coverage, editorial focus, and public discourse in Turkey.
• Students: Utilising a real-world dataset for educational projects in data analysis, linguistics, and machine learning.

Dataset Name Suggestions

• Sabah Newspaper Articles: 2017-2021
• Turkish News Article Dataset (Sabah)
• Turkish Newspaper Content 2017-2021
• Sabah Daily Articles Corpus
• Turkish Media Archive

Attributes