Principal investigator: Eva Mayerhöffer. Other team members Jakob Bæk Kristensen (Post.doc), Frederik Henriksen (PhD student) and Tim Ramsland (Research assistant)
The AlterPublics project studies the role of alternative media in the digital public sphere. Alternative news media have become crucial actors in modern societies, and their ideology-driven content presents a critical challenge for journalism, public deliberation and society at large. The project examines how left-, right-wing and anti-system alternative media contribute to the formation of ideological counterpublics vis-à-vis the mainstream public sphere, and hence their potential contribution to societal radicalization and the polarization of public discourse.
As the first project to systematically compare left-, right-wing and anti-system alternative media across four different countries, it provides unique insights about the role of alternative media and how it differs between countries with different traditions of either including or excluding radical voices and viewpoints from the mainstream public sphere.
The project takes a Big Data approach analyzing the entire digital public that forms around alternative news media across various social media platforms including: Facebook, Twitter, Instagram, YouTube, VKontakte, Telegram, Gab, 4Chan and Reddit. As such, the HPC has become an integral part in order to handle the data volume and allow easy cooperation among project researchers. So far, the project has made great use of the HPC for two purposes: First, to construct and analyse large link sharing networks based on more than 800 million social media posts. The computational capacity and computing power of the HPC have been an invaluable component in the construction and analysis of these networks. Second, we also utilize the HPC to train large language models for analysing social media content and news articles to find topical sharing patterns within the alternative news environments.
More specifically, the project mainly makes use of the Ubuntu-based applications (JupyterLab and the Rsync) to develop code, built tools and transfer data to/from our own server. For code development and deployment, we make great use of the JupyterLab application, where we are able to use all packages from the large and rapidly growing Python community. The various machines (with few and many CPUs) allows for a flexible approach to building, testing and running models that our lower computing power of our laptops do not provide. Our project draws on a range of Python packages depending on the project and research aim at hand. E.g. we employ the NetworkX package for network analysis and Top2Vec and BERTopic packages for text analysis.
Broadening and nuancing our understanding of alternative news media and their contribution to the formation of ideological counterpublics, the project contributes significantly to a rapidly expanding research agenda on the role of alternative media in modern mediated democracy. The contribution will be able to guide media professionals, regulators and the public at large on how to tackle the challenges of societal radicalization, polarization and misinformation presented by this type of media.
The project is funded by the Carlsberg Foundation.