100,000 news items categorized into four topics—economy, microsoft, obama, and palestine—collected over an eight-month period from November 2015 to July 2016. The data includes social feedback metrics from Facebook, Google+, and LinkedIn to support predictive modeling and sentiment analysis.
Use Cases
- Train predictive models to estimate social media popularity using the platform-specific feedback counts
- Conduct sentiment analysis on the news item text to correlate tone with engagement metrics on Facebook and LinkedIn
- Develop topic classification models using the four provided categories: economy, microsoft, obama, and palestine
- Evaluate news recommendation systems based on the temporal distribution of items and their social feedback scores
Strengths
- 100,000 news items spanning an 8-month window from November 2015 to July 2016
- Social feedback counts aggregated from three distinct platforms: Facebook, Google+, and LinkedIn
- Four specific topic labels for classification: economy, microsoft, obama, and palestine
- Tailored for evaluative comparisons in predictive analytics and first story detection