Czech News Article Comments with Reactions and Thread Structure

100,000 user comments were collected from Czech news articles published by Seznam Zprávy. The dataset includes user reactions and the hierarchical thread structure of the comments. The author, organization, and specific time period of collection are not provided.

Use Cases

• Analyze sentiment and toxicity in user comments based on the comment text.
• Model the hierarchical structure of conversation threads to study reply patterns and user engagement.
• Train classifiers to predict user reaction types (e.g., likes, dislikes) from comment text features.
• Study linguistic patterns and topic evolution within a Czech-language online news community.

Strengths

• Contains 100,000 comment records, providing a substantial text corpus.
• Includes metadata for user reactions and comment thread structure beyond raw text.

Limitations

• The sample size of 100k comments may be limited for training large language models.
• Data is exclusively from a single Czech news source (Seznam Zprávy), introducing source and geographic bias.
• Unknown collection timeframe limits analysis of temporal trends.

Provenance

Source

Seznam Zprávy (Czech news website).

Collection Method

Scraped or collected user comments from news articles.

Geography

Czech Republic