Abstract: The bottom-up fine-grained urban community discovery based on travel trajectory data can provide an effective basis for the rational allocation and scientific management of urban resources. The existing community detection methods usually analyze the spatial interaction between hotspot regions from the perspective of complex networks, and have some limitations in the expression of the association between Origin-destination (OD) pairs, which can not well reveal the dynamic urban community structure. To solve the above problems, this paper proposes an urban community discovery method that integrates line graphs and trajectory representations. This method exchanges the nodes and edges of the traditional topological graphs, establishes a new trajectory interactive topological graph with OD pair features as nodes and OD flow start and end points as edges. Furthermore, a visual analytics system is designed and implemented for urban community discovery from the line graph perspective, which helps users explore fine-grained urban sub-regions that can reflect people's real activity space, and provides reference for urban management.