As a novel type of real-time social networking service, microblogginghas already become ubiquitous and an irreplaceabletool. Tracking in the pulse of retweeting propagation isimportant and meaningful. In this paper, we investigate howinformation propagation in a specific microblogging platformevolves to identify relevant patterns and understanddynamic attributes of information propagation and the underlyingsociological motivations. More specifically, basedon the node-link diagram, we propose three efficient strategiesto map the multiple attributes of information propagationgraph to appropriate visual elements. For revealingthe dynamic attributes, we propose two models: the depthvaryingand the time-varying parallel data model to illustratethe temporal evolution efficiently. We also present anovel method by combining the traditional scatter plot withHough transformation to represent the distribution of propagationinstances and trace the propagation speeds. We integrateour methods to a visual mining tool and develop severalinteractive features. We demonstrate how our approachesimprove the understanding of the propagation graph froma visual perspective by employing propagation datasets collectedfrom Sina Weibo, the largest microblogging serviceprovider in mainland China. Meanwhile, this visual miningtool has been evaluated by data analysts and successfullyused in Sina Corporation as a helpful assistant to them.