In this paper, a suite of software components for building visual data analysis system is introduced. Since the rapid developments in web and web-related technology, the need of component-based software is increasing rapidly. Software can be constructed by assembling modular piece-part components. The web-based interfaces present new opportunities for better data analysis methods.

For statistical analysis four important components to web technology are: Browsers, http protocol, CGI-bin programs and New programming languages such as Java, VRML, etc. The web technology has made client-server computing easy. Since the browsers can access to powerful servers, so it is possible to provide the users with a portal into a rich computing environment. However, there are some limitation on web technology, the main problem is it is overloaded. A second problem is security and the related privacy concerns.

This research focuses on building a presentation layer for visual data analysis. The idea is building a set of linked interactive views, each using unique graphical representation or visual metaphor for representing data. The views are packaged as software components and are easily integrated into web browsers to build useful applications. The views work together to provide a multi-view, visual data analysis environment. There are several features of the views: High interactivity, View linking, Filtering, Colorful and Scalable.

In this paper, there are 10 production views in the initial release of the Visual Insights ActiveX Components: Lines, Pies, Barchart, Histogram, Data sheet, Counts, Scatterplot, Parabox, Data Constellation and Time table. Each is described in detail with a figure.

Next, the software architecture is described. Then supported knowledge discovery tasks are addressed. There are even a brief history of this research.

In discussion, it is expected the rapid acceptance of the web and the emergence of component specifications will solve their dilemma of narrowly used Vz library they developed. They will be able to distributed the components widely.

No website is mentioned in this paper.