An Extended Framework for Visualizing the Data from Both Local Databases and Semantic Web Databases Wei Shi, Bin Piao and Yuzuru Tanaka Meme Media Laboratory, Hokkaido University West8. North13, Kita-ku, Sapporo Hokkaido, Japan {shiwei, piaobin, tanaka}@meme.hokudai.ac.jp ABSTRACT Many modern visualization frameworks support users to declaratively define charts by their logical specifications. But these systems do not support the declarative definition of nonstandard charts. To solve this problem, we proposed a new visualization framework. Users can use it to declaratively define both standard charts and nonstandard charts from their logical specifications. Such a specification is described by one or more pairs of Data View Schemata (DVSs) and Chart View Schemata (CVSs). Each DVS is used for retrieving and processing data. Each CVS is used for defining the appearance of each chart component and defining how to use graphical objects to represent data. Because the rapid development of Semantic Web, we further extend our visualization framework to support users to exploratorily search Semantic Web through a node in the DVS. By manipulating the DVS, users can define a view of Semantic Web Resources. Users can also further manipulate this generated DVS to integrate the data from local databases and those from Semantic Web, and then visualize the integrated data with either a standard or a nonstandard chart. KEYWORDS Data Visualization, Hierarchical Data Schema, Hierarchical Visualization Schema, Exploratory Search of Semantic Web Resources, View of Semantic Web Resources, Data Set Integration 1 INTRODUCTION Database visualization is a widely-used technology which supports users to represent data by visual properties of a set of graphical objects in a chart. By observing such kinds of charts, users may obtain the information which may be difficult to obtain from the original data. Because the development of data gathering technologies, multivariate data sets become popular. How to create user-customized charts to effectively visualize such kind of data sets becomes necessary. We classify all the charts into two categories, i.e., standard charts and nonstandard charts. A basic standard chart denotes such a chart composed using four types of basic graphical objects, i.e., “Point”, “Line”, “Area” and “Interval”. A standard chart is either a basic standard chart, or a chart composed with more than one standard charts by geometrically arranging them or embedding some of them into another in a nested way. A chart that is not a standard chart as defined above is a nonstandard chart. In our research [1], we proposed a new visualization framework. Users can use it to declaratively define various charts by their logical specifications. By using our framework, users can separately retrieve and process the data, and define how to visualize the result data by manipulating two kinds of tree-structured schemata. By manipulating these two schemata, users can define both standard and nonstandard charts. When someone reads a chart, he may want to know some further information related to this chart. The data from the local database which is used to define this chart may not include their desired information. He may often need to search the web to further gather his desired information. ISBN: 978-1-941968-02-4 ©2014 SDIWC 76 Proceedings of the International Conference on Computer Graphics, Multimedia and Image Processing, Kuala Lumpur, Malaysia, 2014