Deting Connectivity Patterns in Nordic Twittersphere by Cluster Analysistec

Abstract We analyze Nordic social media users by clustering them based on their connections on Twitter. The data consists of 15,794 users in the five Nordic countries: Finland, Sweden, Norway, Denmark, and Iceland. We first create an undirected graph from the friendship relations (mutually following each other), then divide the graph into five clusters using a recent M-algorithm, and finally compare the results to users’ locations. The results demonstrate that the users are strongly clustered according to their home country. There is surprisingly little interaction across the countries despite the fact that they are, except for Iceland, physically close to each other and have cultural and linguistic similarities. The main language of the four countries belongs to the Germanic languages, while Finnish is typologically distinct. We further explore content from users in each country, analyzing its alignment with connectivity patterns. Our findings reveal a discrepancy between user-generated content similarity in the Nordic region and their connectivity patterns.