Since the development of Big Data methods and their arrival in the social sciences, a number of highly ambitious initiatives have been launched with the aim of changing the way history is researched.

However, the deployment of these new approaches comes up against a great deal of reluctance on the part of historians, who, faced with the difficulties of interdisciplinary dialogue, are often skeptical about the very aims of a collaboration in which they fear, sometimes justifiably, that they will be dispossessed of material they feel they no longer control when it is transformed for inclusion in databases. Both transformation and analysis procedures give rise to profound methodological and even epistemological doubts, especially as the tools used are often innovative and have not benefited from extensive feedback.

The aim is to develop large-scale historical databases by applying data mining methods, in particular around the analysis of relationship networks. Development will be carried out using an iterative approach to the exploration process, based on the appropriation by users of the procedures and tools mobilized, as well as of the results of the analyses. To achieve this, the emphasis will be on making algorithms explicable, and on progressive data analysis and human-machine interaction.

Contacts : Jean-Daniel Fekete | Christophe Prieur