Abstract: A biographical visualization is a graphical description of a person's life to identify key information and gain insights from massive biographical data through visualization and visual analytics. However, automatically generating visualization from biographical text is challenging. To tackle this problem, we introduce BioVistory, an end-to-end visualization generation approach to supporting storytelling with biographical text data. BioVistory uses biographical text and auxiliary information as data input, and visualizes the extracted events, perspectives, and stages by visual metaphors. Users can interactively explore at different semantic levels. We evaluated BioVistory through case studies and a user study, showing that BioVistory can generate high-quality biographical visualizations and support data-driven visual storytelling as well as user-driven reading experience.