Bridging the Divide: Detecting and Disrupting Political Echo Chambers on Social Media

Increasing political polarisation in online discourse, particularly on platforms like X (previously known as Twitter), threatens democratic discourse and critical thinking in politics. This paper explores political echo chambers: ideologically closed online communities that tend to promote confirmation bias and minimise exposure to counter-arguments and opposing ideologies, and misrepresentation of facts and misinformation. In this paper, a social network graph is created in order to provide a representation of online political discourse and opinion on informational platforms like X (Twitter). Nodes denote individual online users, and edges denote online discourse that is facilitated by hashtags. Utilising the Louvain algorithm, with a cosine-similarity filter, the algorithm detects online political affiliations that do not easily align with established notions of left and right ideologies. This algorithm places a premium on the so-called "bridging tweets", or tweets that tend to attract a diverse range of ideologically balanced online engagement, and uses complex natural language processing algorithms in detecting online sentiment with a credibility-weighted influence algorithm that emphasises ideologically balanced online engagement instead of virality.