Welcome Guest. | Log In| Register | Membership Benefits

Intelligent Enterprise

Better Insight for Business Decisions

Home
Digital Library
Events
RSS | Newsletters
Webcasts

May 28, 2002

Vision of Intelligence

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

By Dan Sullivan

Analytics is the art of making sense of masses of data. CRM systems, data warehouses, and corporate portals and intranets provide plenty of raw materials to work with, but the analytic artisan needs a variety of tools and techniques to shape and mold this data into actionable knowledge.

Basic bar graphs and charts have been available almost as long as spreadsheets, but newer visualization models are gaining a foothold in traditional numeric applications as well as in hypertext, relational data, and unstructured data applications. The benefits of visualization vary by application and data type. In some cases, these techniques allow you to quickly compare the relative scale of data and detect patterns. In others, they provide a richer context for understanding data. And in still other cases, visualization techniques let you navigate related but distributed and unstructured data. Regardless of the particular technique, these visualization methods can reduce the time it takes to turn data into information, which in turn reduces decision-making time.

In this article, I'll describe the four most common types of visualization applications, starting with the most familiar: numeric measures.

Executive Summary

Dan Sullivan

Visualization techniques that were once used only for numeric data are now used for relational data, hypertext, and distributed, unstructured data. Several different models are emerging as vendors embrace visualization tools as a method of managing growing volumes of data.

Visualizing Familiar Territory

Among the most common business applications of visualization is multidimensional analysis. Star schemas and related models are widely used in business intelligence (BI) applications and lend themselves to visualization techniques because dimension-based structures provide an easily understood basis for comparisons. Visual Insights' eBizinsights, for example, builds a SQL Server-based data "webhouse" from Web logs and provides a library of predefined reports for analyzing customer behavior and product performance. (See Figure 1.)

This type of report is common but visualization isn't just about volumes and conventional BI measures. Even in the realm of conventional data mart applications such as e-commerce Web-site performance, other visualization techniques can come into play. For example, the Visual Path Analysis tool included in the eBizinsights suite, shown in Figure 2, provides a visual metaphor for analyzing Web-path traffic through a site, making it easier to spot high drop-off pages as well as the most effective paths through a site.

Modeling movement through a Web site by using a map of pages and links is a better visual metaphor than presenting a bar chart displaying a number of page views. The map of pages and links provides cues about the context of the measure instead of just the measure itself, as is the case with bar charts. This context in turn provides more information, such as where page hits originated, captured in a single visual metaphor.

Nothing is inherently "wrong" with bar charts. As with any visualization metaphors, the trick is finding the right one for the problem at hand. Miner3D's Web navigation tool, for example, uses a bar chart model to map page ranking, page size, and domains (for example, .com, .edu, and .gov) into a block-based visual display. The value of the bar chart here is that it lets users quickly assess three attributes of a search hit that are normally displayed as three distinct text elements in a search results listing.

Path analysis and block displays of search results are just two ways of visualizing Web-based information. Hypertext links lend themselves to visual navigation aids as well.

Navigating Hypertext

We've all run into problems finding information on the Web or an intranet because of low precision and recall keyword searches. Taxonomies, such as Yahoo and the Google Directory, provide one way of organizing information so you can find what you're looking for by browsing in context. This approach is useful for finding content on a general topic but it doesn't help you understand the context of a particular document.

For example, a corporate taxonomy might have a category on quality control processes that includes links to best practices used across the organization. What the taxonomy doesn't show is the context of the individual documents; for example, information about an extrusion process might be linked to equipment manuals, material safety data sheets, and other manufacturing process documentation.

Inxight Software Inc.'s Star Tree products, including a viewer and SDK, use a visualization technique called "focus + context" to highlight a particular area of a hyperlinked map (the focus) while displaying a large amount of closely linked items (the context). The focus + context approach quickly gives you a sense of where a hypertext document fits in a larger collection.