Search

Mobilisées dans de nombreuses initiatives citoyennes, les interfaces numériques se retrouvent prises dans une « injonction à la transparence » aux enjeux ambigus. Cette contribution met ainsi en évidence trois paradoxes : 1. Le fait que les interfaces numériques reposent sur une obfuscation ontologique des couches programmatiques. 2. La tendance des projets politiques de transparence à invisibiliser les controverses grâce à l’utilisation rhétorique des interfaces et de leurs données. 3. La transformation du paysage médiatique de la transparence en une tension entre divulgation, mise en visibilité de processus, et création d’expériences de simulation renouvelant potentiellement les capacités de narration des citoyens.

Cet article explore la manière dont les lecteurs reconnaissent leurs identités personnelles, représentées à travers des visualisations de données. La reconnaissance est étudiée à partir de trois définitions, développées par le philosophe Paul Ricœur : l’identification à la visualisation, la reconnaissance de quelqu’un dans la visualisation et la reconnaissance mutuelle se produisant entre les lecteurs. Alors que ces notions étaient initialement utilisées pour étudier le rôle du lecteur de livres, deux concepts supplémentaires complètent le passage à la visualisation des données : si l’identité numérique reste le passeport actuel des actions humaines, la promesse est la réflexion intime qui projette les lecteurs vers leur propre avenir. Cet article porte sur le sens délicat de l’identité numérique et sur la manière de la représenter, selon cette structure : De l’identité personnelle au media est une introduction historique à la reconnaissance de soi, La visualisation de données pour représenter des identités met l’accent sur la représentation visuelle et Le parcours de la reconnaissance brise la reconnaissance de soi à travers les cinq concepts mentionnés ci-dessus, juste avant la conclusion.

International audience

This visualisation represents the research activity of the Digital Humanities Lab through publications and co-authorship. Vertical disposition is ordered by time: each layer is a different year of publications, from the lab’s foundation to nowadays. The layers displays the collaboration networks: two researchers are linked if they published together. The vertical trajectories represent the activity of a researcher through the time. The authors position is fix in the space. As consequence, the trajectories become a linear representation of collaborations continuity. The laboratory is here transformed in a geometrical structure which evolves in time despite the members instability.

Network visualizations are the most complex visualizations possible, but sometimes they are not capable of describing system-complexity. Even if they are the most widely employed visualization techniques, they still have limitations. Indeed a) their relations are not sufficient to analyse complexity and b) networks do not distinguish between qualitative differences of represented entities. Starting from the actual network model, how could one manipulate this visualization to improve complexity comprehension? In this paper, we propose a solution called t rajectory. The trajectory has two major points of difference compared to the network: the trajectory a) represents not only distances, but also durations, and b) it displays kinetic entities according to their evolution with time. The discourse is articulated around these four points. Considering that networks are tools widely used by digital humanists, we propose a new language to improve the quality of represented data: a new network based on a vertical timeline. Complexification of the network visualization is not just a new language, but also a tool that would give the field of Digital Humanities the most complex of all possible visualizations.

This thesis investigates how algorithmic transparency is performed in French public sector organisations. I analyse the practices, methods and performative style of public disclosures in an ethnography initiated at Etalab — the French Open Data task force — and conducted in 2018. The study starts by considering recent controversies about unfair administrative algorithms. This research shows that to be effective, calling for algorithmic transparency requires the staging of identities, issues and algorithms. I describe how information about algorithms is disclosed through the mise en scène of citizens’ motivations, the placing of controversial requests on public bodies, and a regulatory framework redefining administrative procedures as “algorithms”. In a second empirical chapter, I unpack the dispute about unfair calculations of the housing tax. This dispute provides an opportunity to understand how the performance of transparency is purposefully planned by Etalab and the General Directorate of Public Finance. When these two organisations realise that full accountability of the housing tax algorithm is impossible, they set the boundaries of what should be made public. I posit that to be performed, algorithmic transparency requires the negotiation of its limits and the scripting of disclosures. I then study how the housing tax algorithm was disclosed in practice. Since full accountability is not attainable, algorithmic transparency is not longer defined as a bureaucratic duty but performed through a proclaimed exemplarity. Disclosures provide occasions for actors to brand transparency as an honourable achievement, but one that is disconnected from accountability requirements. On this basis, I develop the argument that algorithmic transparency is best understood, not as an accountability device, but as a political force shaping new narratives about public sector digitalisation. The performance of algorithmic transparency serves as an incentive to reorder public services, and, an attempt to refresh the technologies supporting administrative action in the public sector.

Self-recognition is an intimate act performed by people. Inspired by Paul Ricoeur, we reflect upon the action of self-recognition, especially when data visualization represents the observer itself. Along the article, the reader is invited to think about this specific relationship through concepts like the personal identity stored in information systems, the truthfulness at the core of self-recognition, and the mutual-recognition among community members. In the context of highly interdisciplinary research, we unveil two protagonists in data visualization: the designer and the observer - the designer as the creator and the observer as the viewer of a visualization. This article deals with some theoretical aspects behind data visualization, a discipline more complex than normally expected. We believe that data visualization deserves a conceptual framework, and this investigation pursues this intention. For this reason, we look at the designer as not just a technician in the visualization production, but as a contemporary ethnologist - the designer as a professional working in a social environment to comprehend the context and formulate a specific inquiry with the help of appropriate visual languages.