Study On Methods For Multiple Format Information Visualization Facing Data Space

As continuous development of science and technology, information visualization is becoming more and more popular, and is in touch with people's daily lives closer and closer.Information visualization is a technology of human-computer interaction with study content of large-scale non-numeric data resources, and is designed to help people understand and analyze data. Information visualization's development to the present, has generated a lot of visualization techniques which have been applied to many fields, such as finance, business,literature, and other aspects of abstraction. However, hardly anyone study on information visualization in dataspace field.Visualization of data space is divided into two areas, the first is visualization of relationship between visual objects, and the second is visualization of specific object instance.As distributed data and heterogeneous data in dataspace, it is not easy to figure out semantic information of them so that we can't build their effective schema. These features bring many challenges to its visualization, and it is difficult to use existing visualization methods. For relationship visualization, we often use graph layout algorithms, but traditional graph layout algorithms can't make use of data attributes and data patterns; for object visualization,traditional visualization methods need integrate heterogeneous data which step takes a long time in first place and then specify the mapping between data attributes and visualization attributes manually.In this paper, we use two methods to resolve problems of information visualization in dataspace. For association visualization, we extend semantic meaning of traditional Triple model so that it can describe class; we introduce virtual nodes to FR, the force directed graph layout algorithm, so that we can distinguish class from object with four different forces. For object visualization, we first formalize visualization tasks, then convert the problem of mapping between data attributes and visual attributes into the problem of schema matching,finally introduce several heuristic rules to handle this schema matching problem, so that we can skip the step of data integration which takes a long time.Finally, we design and implement iVis, an information visualization module applied to iDataSpace system. Then, we give some examples of real usage scenarios, and the results show that visualization methods in this paper can resolve the problems of information visualization in dataspace effectively.