MUTADATA

Big data is in fashion. In the media as in the public debate, the term is brandished as a banner of innovation. This enthusiasm is accompanied by the highlighting of new expertise (data analysts, data scientist, data miner, data protection officer, etc.) to take advantage of this trend and sell know-how to political institutions in particular. The "data workers", inspired by the American model, claim to be able to "predict" voter behavior through big data and thus "scientifically" affect the results of an election, indicating that they could make all the difference in the final result by enabling gains of 1 to 2 points. Some even argued that their services in electoral big data were a way to tackle abstention. This would mean that the right data handling and skilled data professionals would be enough to win an election. To simplify in these terms, the discourse of data workers seems to be rather a fantasy or a deception. However, the rapid creation of specialized agencies in data management attests to the success of this rhetoric of innovation through data. This research program questions the professionalization of a new political expertise under the prism of big data. Taking a look at providers specializing in database collection, management and analysis allows us to examine the fluctuations of partisan boundaries and the reconfigurations of political work, through the analysis of a composite and hybrid professional space torn between professionalisation, expertise and activism. This project aims more precisely to study the modalities of setting up a professional group and to analyze the effects of these professional recompositions on political work. For each of these areas, the focus will be on both the actors and the socio-technical devices they mobilize. The focus will be on the diffusion and transnational circulation of professional standards and the belief in the effectiveness of web-campaigns that build and legitimize these digital political professions. Combining an approach of sociology of professions, economic sociology and political sociology of parties, we will conduct a comparative analysis of French and American professional spaces, including through an immersion in two data analysis electoral agencies. We will also conduct surveys by semi-directive interviews with American and French "data workers", observations of work practices in other agencies or co-working spaces, mapping of networks of actors, as well as a questionnaire survey of professionals in digital policy communication in general. The contribution of our fellow computer scientists and mathematicians will "open the black box of data analytics" and analyze the fabric of predictive algorithms. This study will provide information on more general phenomena: professionalization and political communication, electoral law, the protection of personal data, fluctuations in partisan borders but also the circulation of standards on an international scale. This project proposes to make an empirically supported contribution to debates and controversies relating to institutional model transfers of change. Key words: big data, political work, professionalisation, election, political communication, United States.