SpacePhaser: Phase Space Embedding Visual Analytics

In the paper, we focus on the statistical detection and understanding of signaling behavior via various media outlets that are associated with a violent hate crime. Following a body of theory that identifies the links between hate rhetoric with violence, we explore textual patterns of hate speech in traditional and non-traditional media outlets, including television news shows, social media, political blogs, and political forums, to detect an association with hate crime events per time and place. Applying the dynamical system theory, we can characterize the dynamic behaviors of essential keywords. Grouping keyword with similar temporal signatures enables us to highlight important information/topics corresponding to emerging events. We also introduce SpacePhaser, an interactive visual tool for textual pattern analysis. To show the usefulness of SpacePhaser, we demonstrate the results of its applications on various real-world datasets.