Research Seminar: Discovering Social Conflicts based on Social Media Data

Social events such as massive protests are likely to turn into social conflicts by some triggers, leading to critical issues for urban governance and society stability. A quantitative understanding of their spatiotemporal patterns, including that how a protest is triggered, how a protest is propagating over time and over geography, what factors may influence the location of a protest occurring, etc., could help urban governors make better decisions for protecting citizens and make effective urban sustainable policies. Social media has been a popular venue for sharing such information before, during and after the social events. By analyzing not only the textual information but the tagged geoinformation involved in a massive social media dataset related to the events, one can investigate the key components of the events. Hence, this work focuses on applying a data-driven framework to a geo-tagged twitter dataset collected for the 2019-2020 Hong Kong protests as a case study. We explore the features of the data through the lens of a set of quantitative analyses such as sentiment analysis, time series analysis and spatial analysis, trying to capture the underlying spatiotemporal patterns of the protests and their violent conflicts.

Poster