• data-driven analysis
• Dash
• social network analysis
• Twitter
• sentiment analysis
• filter bubble

Nowadays social media have acquired an important role in daily news consumption, but they may also become a venue of selective exposure.
The personalization of new content and Internet applications has represented an important digital progress within the past years. In relation to online news consumption the concepts of Filter Bubble and Echo Chambers emerged in this scenario, gaining a focal point both in science and in popular press.
Various factor such as homophily, information overload, congeniality bias, and filter bubbles may lead people to expose themselves to congenial information, consuming only information that align with their beliefs and excluding them from the contradicting one.

This framework tries to figure out whether or not if in the Italian Twittersphere there is the presence of Filter Bubbles, focusing not on the conventional way of conceiving the concept of filter bubble, but revisiting it, analyzing both the content of tweets, but seeing in particular if the behavior and the way of using the platform are more or less uniform between a user and his circle of friends.
This is done through a data-driven analysis of Italian users: with a focus on hashtags, topics, sentiment analysis and community discovery, it was possible to find evidence that to an extent, Twitter is a place where Filter Bubbles are present, supporting findings of some earlier academic studies.
An other main feature of this thesis is the realization of a Dashboard in order to display the statistical, content and sentiment analysis made on the download data.