A Big-Data Approach to Contemporary French Politics

Abstract

Mixing the methods of machine learning and those of French cultural studies, this article explores recent trends in French politics. We analyze a large data set from Twitter (over a hundred and fifty million tweets) in order to reveal notable currents in discourse around the Gilets Jaunes and the French presidency (with particular focus on Emmanuel Macron and Marine Le Pen). This article delineates our methods and places this data analysis in a broader cultural and political context, concentrating primarily on major trends in French politics from 2017 to late 2019. We argue that our interdisciplinary approach reveals elements of French political discourse that would otherwise remain obscure, especially regarding the rise and fall of the Gilets Jaunes movement on Twitter.

Keywords:

• Twitter
• massive data
• French politics
• Gilets Jaunes
• Macron
• Le Pen

Notes

Notes

1 We would like to acknowledge the work of the MDI data blending and technical team, as well as the following students: K. Kawintiranon, Alexander Oddo, Amerisa Kyriazis, and Samantha Melconian. We would also like to acknowledge the funding support of MDI and the NSF grants #1934925 and #1934494.

2 See “Mobilisation, blessures, arrestations… Un an de Gilets Jaunes en chiffres.” CNEWS. 15 Nov. 2019. https://www.cnews.fr/france/2019-11-15/mobilisation-blessures-arrestations-un-de-gilets-jaunes-en-chiffres-840398. Accessed 4 Aug. 2020.

3 Supporters of the Gilets Jaunes have chronicled on the web and on Twitter injuries suffered by protesters, often in gruesome detail. See Le Mur Jaune, a photographic census of alleged injuries incurred during the Gilets Jaunes demonstrations. Last updated 19 Dec. 2019. https://www.lemurjaune.fr. Accessed 4 Aug. 2020.

4 See Damien Leloup, “Malgré les lois, l’Etat a abandonné aux réseaux sociaux l’arbitrage de la liberté d’expression.” Le Monde. 19 Feb. 2020. https://www.lemonde.fr/pixels/article/2020/02/19/malgre-les-lois-l-etat-a-abandonne-aux-reseaux-sociaux-l-arbitrage-de-la-liberte-d-expression_6030122_4408996.html. Accessed 4 Aug. 2020.

5 See “Décision n° 2020-801 DC du 18 juin 2020 Loi visant à lutter contre les contenus haineux sur internet.” 18 June 2020. https://www.conseil-constitutionnel.fr/decision/2020/2020801DC.htm. Accessed 4 Aug. 2020.

6 See Pierre Vermeren, La France qui déclasse: Les Gilets jaunes, une jacquerie au XXIe siècle. Tallandier, 2019.

7 See Jim Tankersley, “U.S. Will Impose Tariffs on French Goods in Response to Tech Tax.” New York Times. 10 July 2020. https://nyti.ms/2BQkGHX. Accessed 11 July 2020.

8 See “Social Media Usage in France.” Statista. 11 Aug. 2020. https://www.statista.com/topics/6278/social-media-usage-in-france. Accessed 14 Aug. 2020.

9 See, for instance, Samuel C. Woolley and Philip N. Howard, Computational Propaganda. Oxford UP, 2019; as well as Yochai Benkler, et al., Network Propaganda. Oxford UP, 2018.

10 https://www.frenchpolitics.mdi.georgetown.edu.

11 Christophe Guilluy, Le Crépuscule de la France d’en haut. Paris, Flammarion, 2017.

12 See, for example, Michel Onfray, Grandeur du petit peuple. Albin Michel, 2020; François-Bernard Huyghe, et al., Dans la tête des Gilets Jaunes. Versailles, V.A. Editions, 2018 ; Gérard Noiriel, Les Gilets Jaunes à la lumière de l’histoire. La Tour d’Aigues, Le Monde / L’Aube, 2019 ; Hubert Artus, Gilets Jaunes: Pour un nouvel horizon. Vauvert, Au diable vauvert, 2019 ; and Gilets Jaunes: Jacquerie ou révolution. Montreuil, Le Temps des Cerises, 2019.

13 This project is hosted under the umbrella of Georgetown University’s Massive Data Institute (MDI), which has a research thrust focused on understanding the impact of different forms of media on public opinion. This research thrust is under the direction of Lisa Singh.

14 On the latter, see Andrew Sobanet, Introduction. Revisioning French Culture, edited by Sobanet. Liverpool UP, 2019, pp. 3–5. The term “massive data” (also often referred to as “big” data) has no standard definition: it simply applies to a data set that is too large to be processed on a standard personal computer. Depending upon the discipline, this can range from hundreds of gigabytes to hundreds of petabytes.

Additional information

Notes on contributors

Andrew Sobanet

Andrew Sobanet is a Professor in the Department of French and Francophone Studies at Georgetown University. He is the author of Generation Stalin: French Writers, the Fatherland, and the Cult of Personality and the editor of Revisioning French Culture.

Lisa Singh

Lisa Singh is a Professor in the Department of Computer Science and a research professor in the Massive Data Institute at Georgetown University. She is the co-author of Words that Matter: How the news and social media shaped the 2016 Presidential Election.