Cartia Organizes Massive Amounts of Information
onto Interactive Topographical Maps
ACTION ITEMS
Cartia, Inc. has announced the immediate availability of ThemeScape 1.0, a new application for managing high volumes of information. Using a process known as Enterprise Information Mapping (EIM), ThemeScape automatically harvests massive amounts of unstructured text, extracts the major themes and topics, and creates an interactive map of the information. The topographical maps show a big-picture view of the content within large quantities of documents, web pages, e-mail, and other sources.
"Very seldom do I use the term 'breakthrough technology.' This is one of those times," said Jothy Rosenberg, president and CEO of NovaSoft Systems, Inc. "A ThemeScape map reveals tremendous amounts of information without ever having to read a document. When you see everything in context, you're able to draw conclusions much faster."
"It's like looking at information from 30,000 feet," noted Greg Clark, a knowledge management consultant and beta user at British Petroleum. "You can see the patterns and relationships that emerge when you combine information from thousands of different documents. ThemeScape has given us insights we may have missed entirely."
The map is the information
The topographical landscapes produced by ThemeScape are not just maps of documents. They are the information within those documents. Major themes rise from the surface, much like mountain peaks, indicating a concentration of documents around a given topic. The distance between the peaks indicates how closely different themes are related. For example, peaks labeled "Internet" and "E-Commerce" might appear in close proximity, but relatively distant from a peak labeled "Keyboards." These visual cues rapidly accelerate a user's ability to understand the context of information.
ThemeScape generates the maps and assigns topic labels automatically, a key requirement when organizing high volumes of incoming information. There is no manual classification or taxonomy, and no tagging of individual documents. Because the system actually extracts content, the maps and labels are automatically tailored to the information being harvested.
"In some ways, you can think of ThemeScape as an old solution applied to a new problem," said Mark Goros, president and CEO for Cartia. "Maps have been commonplace for centuries and have always been used to show information in context. Everybody understands maps, and the model is perfect for textual information. Most users pick up the basics of ThemeScape in a few minutes."
Working with an information map
A ThemeScape map presents an overview of what is contained in a given set of information. In many cases, this high-level view can trigger new insights. For example, a map showing thousands of customer comments would immediately reveal patterns of complaints and suggestions, as well as the relationships between them. These types of insights could be lost entirely if the same customer feedback were assigned to several individuals to read.
The map becomes especially powerful when users begin to interact with it. A good comparison is a simple keyword search. Where conventional search engines often return multiple pages of hit lists, ThemeScape highlights the documents right on the map, visually arranged by topic. This makes it faster to find relevant documents, and may uncover surprises. If a user were researching a competitor, for example, a ThemeScape query on the company name would instantly reveal major areas of activity -- perhaps where it was least expected. Users acquire very important information without ever opening a document.
Roots in military intelligence
The underlying technology of ThemeScape has been in development for more than four years, beginning at the Department of Energy's Pacific Northwest National Laboratory operated by Battelle. With roots in highly classified covert operations, early development was directed by the United States federal government to help discover patterns and trends within hundreds of thousands of intelligence documents. Among other operations, the technology played an instrumental role in analyzing Iraqi troop movements after the Gulf War. In 1996, Cartia secured exclusive rights to the technology and has since redesigned both the mapping engine and user interface for use in commercial applications.
Components, Pricing and Availability
ThemeScape is composed of three primary components:
-ThemeServer is the underlying engine that provides content aggregation, map generation, and information distribution to end users. Each ThemeServer can support multiple ThemePublishers and ThemeReaders.
-ThemePublisher is the control center for ThemeScape. ThemePublisher users can harvest information from local directories, Intranet servers, or Internet web sites. They can customize map appearance, create time-sliced maps based on document dates, and even personalize the application with graphics, user controls, and text specific to their business. ThemePublisher also includes all of the capabilities of ThemeReader.
-ThemeReader is a desktop client that allows users to view and interact with maps created using ThemePublisher. Users can work with the entire information landscape in context, progressively disclose greater detail, search for topics and themes, and view summaries or individual documents.
ThemeServer runs on a Microsoft Windows NT Server. ThemePublisher and ThemeReader are Java applications that run on Windows 95/98 and Windows NT. Pricing varies depending on configuration. An entry system with one ThemeServer supporting two ThemePublishers is approximately $25,000; a system supporting 50 users is approximately $75,000.
ThemeScape 1.0 is immediately available and currently in use at several major corporations, among them British Petroleum, Philips Electronics, Ford and Texaco.
About Cartia
Cartia specializes in Enterprise Information Mapping (EIM), a process that extracts the content of unstructured text and organizes it into interactive maps of information. The company's flagship product, ThemeScape, automatically harvests massive quantities of documents, web pages and other sources; extracts the primary themes and concepts; and creates a topographical map showing the entire information landscape. More information on Cartia can be found at http://www.cartia.com