The Mapping Turk: Human Judgment Driving Algorithms to Visualise Twitter During the 2022 French Presidential Election

Abstract

I had the occasion of visualising the French political Twitter just before the presidential election of 2022, in collaboration with other researchers and the journalists of the French newspaper Le Monde. In this paper, I reflect on this case in an auto-ethnographic style to open the black box of visual network analysis and expose the entangled dialogue between human expertise and computation. I contend that the visualisation’s validity does not root in mechanical objectivity because human judgement was involved at multiple levels, even though that work is not visible in the produced image itself. Like the proverbial “mechanical Turk”, a 18th century chess-playing automaton actually hiding a human player, this big data visualisation hides a reliance on man-made decisions. I first present the origin and social dynamic of this project, I document the methodology employed, I unpack what the map represents, and I explain how to read it (that section is incidentally relevant to the reader interested in French politics). I then return to the question of human judgment to expose in detail how the map was shaped by a negotiation between the journalists from Le Monde, my own research agenda, our methodological commitments, the algorithms employed, and the constraints imposed by the data themselves.