Social media platforms like X (formerly known as Twitter), Instagram, TikTok, and Facebook are fast-growing microblogging services reporting daily news, social activities, and local and global real-life events. Users can exchange information, links, images, or videos with limited content restrictions. These are converted into important sources for true or false information, fake news, misinformation, and disinformation that can simultaneously benefit or damage society.

In this edition of our research spotlight, we delve into a breakthrough research focus area in COSMOS that solves the issue of analyzing multiple platforms at once: the combination of ‘contextualizing focal structure analysis in social networks (CFSA)’ and ‘utilizing Knowledge Graphs (KG)’. CFSA discovers the influential sets of actors that mobilize campaigns on a given social media platform, while knowledge graphs help fuse different social media platforms. Together, they help in revealing a holistic perspective of who drives campaigns that span multiple social media platforms. 

CFSA was further developed by Mustafa Alassad, Nitin Agarwal, and Lotenna Nwana in their recent paper titled “Uncovering Latent Influential Patterns and Interests on Twitter Using Contextual Focal Structure Analysis Design.” CFSA and KG were combined by Abiola Akinnubi, Mustafa Alassad, Nitin Agarwal, and Ridwan Amure in their paper titled “Identifying Contextualized Focal Structures in Multisource Social Networks by Leveraging Knowledge Graphs.” Both papers are published at the 2023 International Conference on Complex Networks, which took place in November 2023 at Menton Riviera, France.

Focal structures are the parts of a network that contain key sets of individuals that mobilize actions and/or campaigns through social media networks. Focal structures can help inform our understanding of how information travels across multiple platforms.

Discovering focal structures that can promote online social campaigns is important but complex. Unlike influential individuals, focal structures can affect large-scale, complex social processes. They’re difficult to identify, utilizing only the usual traditional social or community algorithms. At COSMOS, we have developed a novel contextual focal structure analysis (CFSA) model to enhance the discovery and interpretability of focal structures by providing context about them through their communication network.

As the Internet has gotten more complex and its users have gotten smarter, so has the complexity of networks increased exponentially. CFSA helps solve this problem by taking context into account when searching for influential users on social media platforms.

CFSA is a promising method that can be applied in any social network to reduce conspiracy theories and misinformation spread, as it helps identify key sets of users that spread this misinformation. We have applied this method to X, Instagram, and YouTube data sets in different fields. For instance, we have successfully pinpointed clusters that disseminate conspiracy theories on YouTube regarding conflicts in the South China Sea; we have also discovered various discussions across different US election campaigns on X that all related to the South China Sea and the broader Indo-Pacific region. These are among the most prominent topics we have targeted, gathering data and applying focal structure analysis to discern prevalent themes and interests.

In identifying contextualized focal structures in multisource social networks, we have then leveraged knowledge graphs. A knowledge graph is a knowledge base that uses a graph-structured model of data to represent and operate on data. Knowledge graphs are often used to store interlinked descriptions of entities—objects, events, users, even abstract concepts—while also encoding semantics and relationships underlying those entities. 

Little research has been conducted that combines knowledge graph models with contextual focal structures that are derived from multi-source data, and, in fact, “Identifying Contextualized Focal Structures in Multisource Social Networks by Leveraging Knowledge Graphs” is, to our knowledge, one of the first research papers to address this issue. The research in this study identifies and contextualizes the focal structures comprising entities, topics, themes, and documents from multiple online social networks. This approach is validated on data from the Indo-Pacific region to discover information campaigns on various social media platforms associated with the region. We plan to develop further and improve this area of research.