Societal Debates

You are here

By analysing discussions on social media and newspaper articles, in this exploratory we study public debates to understand which are the most discussed topics. We can identify themes, following the discussions around them and tracking them through time and space. 

 

STORY 1: Polarised Political Debates

How does people discuss on online social networks? Who are online social network users that take part in political debate? What is the structure of their social network?

In this story we focus on online debates about discussing topics. By using data from online social networks, we analyze people discussions about topics that are relevant for society.

We will investigate who attends dibates, how people discuss, and what are their social relationships compared with those of people of different views.

We try to understand also how politicians discuss and they influence other people.

We made visualization to represent the user social network and topics evolution.

 

British EU referendum data

Past polarised debates from AALTO

Facebook data from CSSLab (awaiting confirmation)

Method Partner SoBigData RI - Integration
GATECloud / YODIE Entity Linking to DBPedia/stance detection USFD Web Service
Mimir USFD Web Service
Digital DNA fingerprinting CNR Web Service
Polarized Tracker CNR  
Egonetworks CNR Service hosted, Download

Dominic Rout - USFD
Kalina Bontcheva - USFD
Giles Greenway - KCL
Aristides Gionis - AALTO
Kiran Garimella - AALTO
Michael Mathioudakis - AALTO
Claudio Lucchese - CNR
Paolo Ferragina - UNIPI
Marco Cornolti - UNIPI
Stefano Cresci - CNR
Maurizio Tesconi - CNR
Valerio Arnaboldi - CNR
Walter Quattrociocchi - IMT

 

STORY 2: Monitoring Topics across Time and Space

What does newspapers talk about? Is it possible to understand when a specific theme becomes relevant for the media? Are there some topics that become interesting for all journalists at the same time?

In this story we analyze german newspaper articles to understand the most relevant topics and monitoring how they are discussed by press. We made a visualization that shows the most frequent topics in a certain period of time.

In our visualization each word is weighed on the basis of its meaning and it has reference on space and time.

Thanks to this study it twill be possible to identify interesting topics automatically, before to start studies each topic.

 

Method Partner SoBigData RI - Integration
Archive subset creation LUH  
GATECloud / YODIE Entity Linking to DBPedia USFD Web Service
TagME Entity Linking to {Wiki,Db}pedia UNIPI Web Service
DEMON/TILES (community discovery in static/dynamic graphs) CNR Download
GeoVis/GeoVA of ST data FRH.IAIS Download
Visualization for Topic Exploration FRH.IGD Web Page

Gerhard Gossen - LUH - Lead, Web archive access
Dominic Rout - USFD - Text mining
Ian Roberts - USFD - Text mining
Aris Gionis - AALTO - Graph mining
Paolo Ferragina - UNIPI - Text mining
Marco Cornolti - UNIPI - Text mining
Thorsten May - FRH - Visualization
Gennady Andrienko - FHR - Visualization, spatial mining
Natalia Andrienko - FHR - Visualization, spatial mining
Giulio Rossetti - CNR - Networks & Network dynamics