Sentiment analysis, topic modelling and social network analysis

Abstract

Extensive anti-government protests were held in Poland in October 2020 related to accumulated social tensions due to COVID-19 pandemic among others. We attempt a question of how understanding communication patterns among protesters could possibly support epidemiological harm reduction campaigns. To do so, we analysed tweets in Polish language with hashtags: #Strajkkobiet representing pro-choice movement woman's strike (227, 125), #zajob representing coronascepticism movement (22, 379), #ProtestRolników for farmers protest (9, 739) and Social Network Analysis (SNA) was applied. Practical recommendations for harm reduction and educational communicates have been revealed: (1)#Strajkkobiet – Due to the modular structure of the network, the organisers or protest leaders would not be able to propagate pro-health communication towards the "teenagers” clusters. (2)#Zajob – Very high network density with very low modularity suggests that such a network is resilient to interventions such as the blocking of single accounts or inhibiting reach ofcontent. (3)#ProtestRolników – Animal breeders' protest communication has the most hierarchical and modular structure of all the networks studied with clear boundaries between communities and opinion leaders who could propagate information campaigns for their followers. © 2022 selection and editorial matter, Katarzyna Kopecka-Piech and Bartłomiej Łódzki;individual chapters, the contributors.