Opinion Dynamics and Political Conflicts in the Media - A Complex Networks Perspective

DATES: Fall 2017 - Fall 2020
WORKSITE: LIP6 / 4 place Jussieu / 75005 Paris
SALARY: 21k€ gross per year (plus eventual teaching responsibilities)

- Université Pierre-et-Marie-Curie
- École doctorale Informatique, Télécommunications et Électronique (ED130)
- Laboratoire d'informatique de Paris 6 (UMR 7606)
- Équipe “Complex Networks” (

- Robin Lamarche-Perrin ([log in to unmask])
- Matthieu Latapy ([log in to unmask])
- Clémence Magnien ([log in to unmask])

This PhD position is founded by the European H2020 ODYCCEUS project “Opinion Dynamics and Cultural Conflicts in European Spaces” that started in January 2017 (

Please send to <[log in to unmask]> before the 28th of May, 2017:
- a complete resume;
- a transcript of your master's grades;
- a summary of your master's thesis;
- a short covering letter (one or two pages) describing in more details what would be your PhD project.
Applications in French and in English are welcome.

Social media and the digitisation of news are having far-reaching effects on the way individuals and communities communicate, organise, and express themselves. Can the information circulating on these platforms be exploited to better understand and analyse the enormous problems facing our contemporary society? Could this help us to better monitor the growing number of social crises due to cultural differences and diverging world-views? Studying the structure of debates in the public sphere requires sophisticated methods for the analysis of information flows between individuals. How is information shaped and broadcasted by mass media? How to describe the way opinions are discussed in social media? Debates are often represented as complex entanglements of such social interactions, embedded in space and time, and displaying a multilevel structure: From individual to institutional discourses; From regional to international matters; From the fast dynamics of media “buzzes” to the slower dynamics of social controversies.

To address these challenging issues, this PhD position aims at developing new methods for the analysis of multidimensional and multilevel networks in social sciences. First, by building on recent work in dynamical graph theory regarding the “link stream” representation of evolving networks [4], which provides a novel and intuitive formalism for the spatio-temporal description of social interactions by focusing on their causal structure (who interacts with whom, when) and concealing for a moment their content (how, why, about what). Second, by integrating recent developments in graph compression [3], which builds on information-theoretical data compression [2], to provide a macroscopic perspective on such interaction structures and thus achieve a global understanding of complex interaction patterns.

The efficiency of the proposed analysis methods will then be evaluated by empirical work on real data, in collaboration with researchers in media studies, in political sciences, and in quantitative geography that are working on the ODYCCEUS project. First, by focusing on the analysis of opinion dynamics in social media, such as Twitter, and looking at the topology and dynamics of interactions related to a particular debate (e.g., climate change, presidential elections, Brexit). Here, a focus will be put on the study of typical interaction patterns between actors of the debate, such as: polarisation, leadership, communitarianism, and solidarity behaviours. Second, by focusing on the analysis of conflicting world views in mass media [1], for example by modelling the news about international conflicts as a dynamical graph between countries and media outlets (citations in articles) or between couples of countries (co-occurrences in articles). The world view of different media outlets will then be compared by graph analysis, thus allowing to exhibit particular behaviours depending on the geographical, political, and cultural interests of the media.

Note that this PhD position does not require any preliminary knowledge of the hereabove mentioned theoretical frameworks (dynamical graph and graph compression), but however requires the capacity to work with abstract formalisms in general, in order to build generic and sound methods within the domain of theoretical computer science. This PhD position also requires a genuine interest for interdisciplinary work, and in particular a curiosity for some of the research questions that arise in social and political sciences about traditional media and social media.

- Developments in graph theory for the analysis of dynamical and multilevel networks: for example, by building on the “link stream” formalism [4] and on information-theoretical graph compression [3].
- Implementation and documentation of algorithms, analysis, and visualisation tools. Integration of these tools to a software that will be developed within the ODYCCEUS project: the “Opinion Observatory”.
- Empirical validation of the theoretical contributions in collaboration with social scientists on several case studies regarding European debates or conflicts (e.g., refugee crisis, Brexit, COP21, European elections).

- Learning about theoretical work in media studies and in political sciences about the theory of agenda-setting to fit the designed methods to the needs of sociological and political analysis.
- Comparing developed methods with those that will be developed in parallel by other partners of the ODYCCEUS project (e.g., spacial interaction models in quantitative geography, agent-based opinion dynamic models in computational sociology).
- Dealing with real, complex, and large-scale data.
- Dissemination of the contributions through technical reports, publications, demonstrations, and networking, in particular during the consortium meetings of the ODYCCEUS project.

[1] Claude Grasland, Robin Lamarche-Perrin, Benjamin Loveluck, and Hugues Pecout. L'agenda géomédiatique international : analyse multidimensionnelle des flux d'actualité. L'Espace Géographique, 45:25–43, 2016.
[2] Robin Lamarche-Perrin, Yves Demazeau, and Jean-Marc Vincent. Building Optimal Macroscopic Representations of Complex Multi-agent Systems. Application to the Spatial and Temporal Analysis of International Relations through News Aggregation. Transactions on Computational Collective Intelligence, XV:1–27, 2014.
[3] Robin Lamarche-Perrin, Tiphaine Viard, and al. A General Framework for Graph Aggregation., CoRR, Forthcoming in 2017.
[4] Tiphaine Viard, Matthieu Latapy, and Clémence Magnien. Computing maximal cliques in link streams. Theoretical Computer Science, 609:245252, 2016.
Robin Lamarche-Perrin

Centre national de la recherche scientifique
Institut des systèmes complexes de Paris Île-de-France
Laboratoire d'informatique de Paris 6

Mail: [log in to unmask]
_____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers ( To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.