The Geometry of Public Issues – GOPI

GOPI introduces a new geometric approach for the modeling and visualization of the trajectory of public problems as they unfold in heterogeneous public spaces. The project objectives are both applied and methodological. We will target three case studies all related to health-related problems in France: (i) the sanitary risks due to occupational or environmental exposure to pesticides (ii) the rise of vaccine hesitancy among the general public, health professionals and activists and (iii) the threats posed by the advent of AI in medicine (computer-powered diagnostics, data-driven healthcare system, etc.). The project will embrace the polyphonic nature of public space and propose to monitor and model the social process of claims-making in public space. An original corpus of speech acts coming from various online sources will then be collected and hosted by the GOPI platform for each issue. We will set up a continuous data collection workflow to capture various digital traces pertaining to the three case studies such as articles published by online media, posts by advocacy groups on Facebook, conversations in public forums, tweets, etc. Original methods for the modeling of public issue dynamics will be implemented in the same platform allowing experts of each case to question the social and semantic structure of each public issue as they unfold in heterogeneous spaces and visualize its transformations. We propose to develop a geometric modeling of public issue dynamics. Word embedding techniques and corpus alignment strategies will be leveraged to position key controversial entities and actors in a continuous space and measure their displacement in the various social arenas they circulate in. We refer to the term “computational hermeneutics” to define our approach, which is both data-intensive and aims at modeling as accurately as possible the argumentative structure developed by actors in their discourses.