Topological analysis has provided insight for understanding high dimensional point cloud that cannot be visualized by conventional methods. The combination of the constructed topology with the visualization methods such as scatter plots or parallel coordinates plots on the raw data has greatly benefit the domain experts in understanding the data-set or simulation process. However, when the dimension increases, the data-set with thousands of points or even millions of points would become sparse or irregularly sampled thus larger data-set is always preferred to show the correct topology or data distribution. However, the visualization of the raw data in the form of scatter plots or parallel coordinates will overwhelm the users when giving too much information in the visualization. At the same time, it will also dramatically slow down the interactive process of querying the data-set to explore different features. Thus a method that reduces the amount of information shown to the user as well as reduces the latency in querying the data-set is preferred. The proposed method combines topological segmentation of the input space with the visualization of the summary information of the data-set. Among all the summary visualization approaches, data aggregation is used and the corresponding histogram is used to show the correct distribution of the data. As a result, the users would not be overwhelmed by the amount of information shown in the visualization results shown as summary scatter plot and summary parallel coordinates as well as understand the data-set through the topology shown as topological spine.