The goal of this project is to better understand the activity in social networks both from a research and a journalistic perspective. It is a challenging task as the access to data is problematic. On the one hand there is an enormous amount of users and posts requiring big data frameworks for handling them and on the other hand, most of the network activity is hidden from the public and as a consequence from any monitoring tool. The scientific part of the project will investigate and model the particular structure of social networks, the spreading of information within them and the possibility to retrieve network activity from an incomplete view of it. It will make use of the recent advances in network analysis, data science and machine learning.

This project is a synergy between researchers and media experts. The scientific results together with the media expertise will be combined to design a program for monitoring the activity of social networks, with unprecedented accuracy. Journalists on the media side will take part in the creation and the test of the program, providing valuable expert feedback on each step of the data processing pipeline. They will also contribute to the development of labelled datasets, a crucial requirement for accurate results using machine learning.

The social network monitoring software will give insights about the users’ activity within the network, enabling fake news detection, the identi cation of polarization and problems in the sharing of information as well as reporting on the user opinions and their evolution over time.