A visually presented schema of a data set can help its user to gain understanding about its contents and use an appropriate vocabulary to build queries or develop applications accessing it. Still, most of Linked Open Data sets are provided without available visual schemas and for data sets of realistic size the schemas may be way too large to be visualized on a suitable diagramming canvas. In this paper we develop a concept of visualization-oriented data schema, involving property ascription point information and propose a method for creating visual presentations of the data schemas, including both the detailed schema visualization and visual schema summarization options. We evaluate the method on 24 prominent data sets from the Linked Open data cloud, where it can obtain legible visualizations of the full schema or at least its fragment.