Key Findings

  • Unsurprisingly, when choosing which story to cover, EU news outlets are more likely to make similar choices if they are in the same country. [REF]
  • Outlets in different EU countries, tend to to cover the same
    stories, if these countries
    • are neighbors
    • vote for each other in the Eurovision song contest
    • have strong trade ties


  • EU countries with "untypical" choices of stories to cover tend to be those outside of the eurozone, and those that joined the EU latest. [REF]
  • There is a strong bias towards covering males in every topic. The bias is lower in entertainment and higher in politics, business and sports. [REF]
  • Articles in entertainment and sport are more readable than articles in business and politics. [REF]
  • It is possible to predict which stories will become popular and find the related keywords. [REF]
  • The content of Twitter can be reliably used to predict the presence of flu in a UK location (as well as its weather state).
  • It is possible to build networks of people and other entities like organizations or places based on their co-occurrence in media content.
  • It is possible to detect memes that are spreading fast through the mediasphere, intended as n-grams (strings of n words) whose frequency suddenly increases in a significant way. [REF]
  • We can track which countries mention which other countries in their media content. [REF]