Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

May 28, 2002

Vision of Intelligence

Visualization technology will soon be an integral part of mainstream strategic business apps

By Dan Sullivan

Continued from Page 1

Rather than laying out each node using a standard 2-dimensional (2D) tree format - which would work well with a network diagram - the focus + context technique of Star Tree gives more space to the node in focus while reducing the amount of space occupied by related, or context, nodes. The benefit of this approach - called a hyperbolic tree - is that the most important information, the focus, is prominent while limiting the space occupied by individual surrounding nodes in order to maximize the context that's displayed. A recent study at Xerox PARC (which spun-off Inxight to commercialize its visualization and language analysis tools; see Resources) found users could browse 62 percent faster with a hyperbolic tree than with a traditional 2D tree layout such as Microsoft Explorer.

Just as visualization for numeric measures help you absorb more information than tabular displays, focus + context methods let you see a hypertext landscape, not just individual points in that area.

But as you'll see, intranets and the Web aren't the only points of information overload in an organization. Relational databases have become prime targets for visualization vendors as well.

Visualizing Relational Data

Visualization Featured in Portals and Related Tools

Visualization tools must necessarily integrate with existing IT applications and data sources. Portal vendors and related technology providers are adapting visualization to enhance their core product offerings and providing these enhancements in different ways.

In the case of Autonomy Corp. PLC, a leading enterprise search and portal provider, the vendor includes a 2D spectral map visualization component in its search tool. Users can quickly move around a map to display information about related content. The visualization tool is tightly integrated with other search features provided in Autonomy's product line.

While Autonomy has chosen to incorporate a visualization tool directly, Semio Corp., a developer of taxonomy generation tools, has partnered with visualization vendors such as TheBrain Technologies and Antarcti.ca Systems Inc. to offer different methods for visualizing taxonomy information.

Anarcti.ca's Visual Net uses a geographical map metaphor to display varying levels of detail. For example, a capital investment application uses a Visual Net map that displays boxes arranged alphabetically by category name with their size varying relative to the amount of capital invested in that area. Points on the map indicate a single investment and visual cues, such as the size of different colored concentric circles to show how much was invested and how recent the deal is. Vendors and developers are adapting visualization techniques to industry- or domain-specific information types as well as a variety of data sources and structures.

Star Tree and other focus + context tools utilize explicit connections among objects so they aren't limited to hypertext links. Relational data can also provide the basis for a visualized map.

Consider, for example, an e-commerce site with a catalog of products. A user could navigate the site by following explicit hypertext links, but this approach could limit visitors to a small number of paths based on a product hierarchy, links to cross-selling items, special promotions, best sellers, or other predefined links.

An alternate method is to let users navigate based upon data relationships, not page relationships. In this scenario, users navigate a web of relations based on product categories. The Sony Music licensing Web site (www.sonymusic.com/licensing), for example, uses ThinkMap Server from Plumb Design Inc. to display the results of keyword searches by artist, track, and subject. Results are displayed in a 2D map showing related topics as well as those specified in the search query.

For example, because songs are categorized with multiple keyword labels, a user could navigate from songs categorized as "Ecology" and "Political Classics" to those categorized as "Political Classics" and "Protest" and then to some indexed as "Protest" and "Crime." Unlike hyperlink based navigation, which is driven by the structure of a Web site, database driven navigation is driven by the organization of the database content.

Corralling Distributed Content

Of course, not all content is in hypertext format or stored in relational databases. One visualization vendor, TheBrain Technologies Corp., has developed an application called BrainEKP designed to work with unstructured and structured information content distributed across multiple platforms. The BrainEKP presents users with a visual map of related content, which the vendor calls "thoughts." These thoughts are structured according to users' needs, and can include, but aren't limited to, predefined hypertext or database relationships.

For example, a sales team responding to a proposal might create a thought with links to the original request for proposal (RFP), working notes, prior responses to the same client and related customers, as well as similar proposals developed by other teams. These items aren't necessarily stored in a single repository. Working notes might be linked from a Lotus Notes database while the RFP is on a file server and similar proposals are kept in a Documentum system. The items in one thought, in turn, can be linked to other thoughts creating a web of contextually related business process instances.

In this example, the organizing principal is the business process that drives proposal development. Visualization is used to navigate to particular steps of the process and display contextually relevant information. Depending on the application configuration, users can add more content creating a de facto collaboration tool for sharing information in a business process specific structure. The value of this type of visualization is that it makes explicit the inherent relationship between information spread across multiple systems.

Visualization Solutions

Visualization tools provide metaphors for understanding large amounts of data. Sometimes we need to assess relative measures (for example, What were the top selling products last month?) Multidimensional analysis tools work well here, but in special cases, such as analyzing users' navigation through a Web site, specialized tools such as path analysis can make relevant information easier to find. In other cases, we need to better understand a single piece of information by grasping its context (for example, "What else is known about quality control problems of this part?"). Here, visualizing hypertext and database-oriented context can provide rapid access to relevant data. Finally, visualization lends itself to complex business processes such as sales development where users need to model a number of different types of information that may or may not have explicitly defined relationships. Tools implementing bar charts, visual path analysis, hyperbolic trees, and user-defined aggregate structures are all available to fill a variety of needs in the intelligent enterprise.

Rate This Article

Comments:

Optional e-mail address:

Dan Sullivan [DSullivan@RedmontCorp.com] is the CTO of Redmont Corp. (www.redmontcorp.com), a consulting firm specializing in the design and development of portal, content management, collaboration, and business intelligence applications. He is the author of Proven Portals (Addison-Wesley, forthcoming) and Document Warehousing and Text Mining (Wiley, 2001).