You are here

Twitter in the French Presidential Elections

The next paper at ECREA 2012 is presented by Jean-Marc Francony, who shifts our attention to the French presidential election and begins by noting the difficult process of shuttling between data analysis and theory-development in the context of political uses of social media. His approach is to consider the election as a media event, and the research builds on some 2 million tweets from the socialist primaries and the French presidential election. Specific live media events during this time were also investigated.

This research considers the Twittersphere as connecting individuals, organisations, and live events. The data include time-specific materials as well as profile information for participants, therefore, and the focus of the investigation was to examine the roles of actors in these events, to identify the audiences for different events, and to explore how Twitter reacts to media events.

This uses a process of stream analysis for specific media events, aligning the Twitter timeline with the timeline of the events. Such work confirms the synchronicity of Twitter and mainstream media – both during specific media events (at the micro level) and over the course of the overall campaign (at the macro level).

This means that the analysis of individual tweets is not sufficient to understand Twitter activity; rather, what the research introduces is the concept of the Twitter flow over a longer amount of time and involving many users, in order to examine the dynamics of Twitter during this time.

Further, network analysis enables the researchers to identify key actors in the process; this also points to the role of such lead users as drivers of the conversation and as editors of information. What emerges from this is that Twitter is part of the media cycle, rather than a separate space; activity there reacts to events during the campaign, and politicians and journalists play an important role during these interactions. But more work, especially also using semantic analysis of tweet contents, still needs to be done.