This paper seeks to understand extreme public transit riders in Beijing using both traditional household survey and emerging new data sources such as Smart Card Data (SCD). We focus on four types of extreme transit behaviors: public transit riders who (1) travel significantly earlier than average riders (the 'early birds'); (2) ride in unusual late hours (the 'night owls'); and (3) commute in excessively long distance (the 'tireless itinerants'); (4) travel over frequently in a day (the 'recurring itinerants). SCD are used to identify the spatiotemporal patterns of these three extreme transit behaviors. In addition, household survey data are employed to supplement the socioeconomic background and provide a tentative profiling of extreme travelers. While the research findings are useful to guide urban governance and planning in Beijing, the methods developed in this paper can be applied to understand travel patterns elsewhere.