Semantic network analysis of Twitter data and their psycho-social implications

To study the language of social networks and its psycho-social implications there is a need for an interdisciplinary approach that combines scientists from the fields of network science, psychology and linguistics. This thesis is a product of the author's work undertaken in two such interdisciplinary projects, Comparing the role of men within prolife and prochoice community in discussion about abortion on Twitter and Analyzing political discourse before and after election using Tweets posted by the candidates for the 2020 US Elections. Both projects used tweets in English language from various users collected over a predefined time span. Twitter data offers various possibilities for interpretation within the context of network science. The focus of this thesis was to study a special kind of networks, the semantic networks, built from the tweet text, hashtags and metadata. More specifically, we detected meaningful communities based on the topics in order to study the language used within the context of those topics. Although in both projects tweets are used to study language on social media and its psycho-social implications, the main goal differs and so does the methodology applied to community detection.