The Assassination of Shinzo Abe

• In Event: Poster Session: Public Opinion and Comparative Politics

Abstract

The study of information, human emotions and public opinion in the digital age faces a litany of technical challenges, including collecting, storing, cleaning, and analyzing data at scale over time. In this study, we propose a systematic approach towards addressing aspects of this “quadrilemma.” We developed a methodology for daily collection of social media posts at the moment they are created, reducing the skewing effects of moderation and deletion in order to obtain an accurate picture of the conversation as it happens. Then, we propose efficient and secure storage procedures for large datasets, before moving on to how to clean and enrich them, which pose multiple hardware and software challenges. We further provide various tools of analysis that can be applied to explore the data. Finally, we execute the proposal across the quadrilemma process on the case of Shinzo Abe’s assassination. The murder of the controversial former prime minister represents Japan’s single largest twitter event (our dataset consists of over 200 million tweets), took place during a national parliamentary election, and was motivated by frustrations related to a cultish religion originating abroad: Shangri-la conditions for an emotional firestorm with public opinion consequences.

Authors

• Matthew Brummer, GRIPS

• CONSTANTIN IOAN BARBU