The phenomenon of online toxicity, particularly during political elections, presents significant challenges for digital platforms, policymakers, and society at large. This thesis investigates the dynamics of toxic interactions on Twitter in the context of the 2022 Italian political elections, focusing on the role of politicians' gender and political affiliation in influencing the nature of online discourse. Through the collection and analysis of a dataset comprising over 500,000 tweets directed at Italian politicians, this study employs statistical methods and machine learning techniques to uncover patterns of online behavior. Key findings indicate that female politicians are disproportionately targeted by toxic tweets, and political alignment significantly affects the volume and intensity of online toxicity. The research further explores the temporal trends of toxic interactions, highlighting spikes in activity correlating with key campaign events. By providing a detailed understanding of online toxicity, this thesis contributes to the development of more effective strategies for mitigating hate speech and fostering a healthier online political dialogue.
Il fenomeno della tossicità online, in particolare durante le elezioni politiche, presenta sfide significative per le piattaforme digitali, per chi le gestisce e per la società nel suo complesso. Questa tesi indaga le dinamiche delle interazioni tossiche su Twitter nel contesto delle elezioni politiche italiane del 2022, concentrandosi sul ruolo del genere dei politici e dell'affiliazione politica nell'influenzare la natura del discorso online. Attraverso la raccolta e l'analisi di un dataset composto da oltre 500.000 tweet diretti ai politici italiani, questo studio utilizza tecniche statistiche e di machine learning per scoprire modelli di comportamento online. Le principali conclusioni indicano che le esponenti politiche di genere femminile sono bersagliate in modo sproporzionato da tweet tossici, e l'affiliazione politica influisce significativamente sul volume e sull'intensità della tossicità online. La ricerca esplora inoltre i trend temporali delle interazioni tossiche, evidenziando picchi di attività correlati agli eventi chiave della campagna elettorale. Fornendo una comprensione dettagliata della tossicità online, questa tesi contribuisce allo sviluppo di strategie più efficaci per mitigare l'odio online e favorire un dialogo politico online più sano.
Toxic Tweets: Unpacking the Gender and Political Dynamics of Online Abuse in Italian Elections
Usubelli, Nick
2023/2024
Abstract
The phenomenon of online toxicity, particularly during political elections, presents significant challenges for digital platforms, policymakers, and society at large. This thesis investigates the dynamics of toxic interactions on Twitter in the context of the 2022 Italian political elections, focusing on the role of politicians' gender and political affiliation in influencing the nature of online discourse. Through the collection and analysis of a dataset comprising over 500,000 tweets directed at Italian politicians, this study employs statistical methods and machine learning techniques to uncover patterns of online behavior. Key findings indicate that female politicians are disproportionately targeted by toxic tweets, and political alignment significantly affects the volume and intensity of online toxicity. The research further explores the temporal trends of toxic interactions, highlighting spikes in activity correlating with key campaign events. By providing a detailed understanding of online toxicity, this thesis contributes to the development of more effective strategies for mitigating hate speech and fostering a healthier online political dialogue.File | Dimensione | Formato | |
---|---|---|---|
Toxic_Tweets_Unpacking_the_Gender_and_Political_Dynamics_of_Online_Abuse_in_Italian_Elections.pdf
Open Access dal 13/03/2025
Dimensione
1.76 MB
Formato
Adobe PDF
|
1.76 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/218100