Investigating Echo-Chambers in Decentralized Social Networks: A Mastodon Case Study

Abstract

There are studies on echo-chambers and homophily in centralized networks, while these studies in decentralized networks are scarce. This work investigates echo-chambers and tightly-knit communities in the decentralized network Mastodon, using multiple methods such as network topology, clustering coefficient, homophily and sentiment analysis. The investigation was conducted with three different datasets, each dataset focusing on one or more hashtags related to a specific topic. We found very few small topic-based tightly-knit communities, and our sentiment analysis showed the existence of relatively few such communities in terms of sentiment alignment about a topic, which is indicates that echo-chambers might not be strongly present in decentralized networks.