Data Science Brown Bag: Towards understanding socio-semantic configurations in networks

Abstract from the speaker: 

When analysing social networks, it is often helpful to consider attributes that are known about the actors, i.e. metadata, to provide further insights into the underlying network dynamics. For example, when considering online public discourse represented by an interaction network of social media users we might not only be interested in the links between the user nodes, but also in additional user information, such as demographics or political preference, and the interplay between such user categories and link formation. One way in which we can integrate metadata in network analysis is to explore the relationship between metadata and what we call the network’s block structure — a division of the network into groups of nodes that are similar in terms of how they are connected to each other and the rest of the network. In the network science literature, the prevalent assumption of an intrinsic connection between metadata and block structure has faced scrutiny in recent years. Nevertheless, a tool for measuring the metadata-structure relationship for the purpose of systematic comparative meta analyses has been notably lacking.

About the speaker: 

Lena Mangold is a PhD researcher in the Computational Social Science Team lead by Camille Roth, at Centre Marc Bloch (Berlin). In the context of the ERC socsemics project, her research focuses on the mathematical description and categorisation of socio-semantic clusters in online public spaces: subgroups of users not only segregated from the wider conversation in terms of social interaction but also separated from (and less exposed to) opinions and content outside of their group.