During the first lockdown, people have largely used the social media platforms to share their opinion and feelings about the Covid-19 and the political choices. Accessing this information can be extremely relevant for the government.
This study presents a new approach for the profiling of social media users and the identification of virtual communities, based on the analysis of communication styles and social behaviors. We use the Emotional Text Mining (ETM) technique for the identification of social media users’ clusters and the categorization of lexical profiles and sentiment. We combine this with social network analysis (SNA) to identify users’ community and to measure their social dynamics. Lastly, we compare the lexical communities with the SNA communities. We show the advantages of our approach by presenting a case study – where we analyze more than 80,000 Twitter messages, written in Italian, during the first step of the Italian lockdown. Through the ETM we detect four communities (enemy within, enemy from abroad, forced confinement, paralysis), each one with its own reaction to Covid-19. Through the SNA we identify the main actors of the spread of information and we find over a 1000 of communities. We discuss the comparison of the lexical and SNA communities that show a partial overlap. Our study advances research on words and networks, offering a new approach for the profiling of social media users, the analysis of Covid-19 representations and the study of the communication-behavior connection. Our findings have important practical implications for the government and the public institutions.
