Abstract: Visualizing sport data is of great importance for sport news and gameplay. Sport data can be categorized into simple one-dimensional statistical data of sports attributes, multi-dimensional data and time-space attribute. Existing work focuses on sport data journalism, professional data analysis and visual analytics system. This paper summarizes four classes of fundamental visualization approaches: spatio-temporal visualization(statistical data, statistical and spatial data, statistical and temporal data, statistical and spatio-temporal data). From the perspective of players, there are two kinds of methods: single player, and multi-players. We present visual analytics systems that emphasize on statistical aspects, movement and constellation, and feature detection, respectively. Finally, we outline future work for sport data visualization.