Visual Search and Information Visualization vis-a-vis Information Retrieval

Broadly speaking, information visualization is a graphical presentation for manipulating information extracted from a larger document corpus or an information database, this ability to represent information a in a graphical user interface enables users to understand and grasp the information faster, recognize and discover meaningful trends, patterns, important information clusters, providing him with more actionable information, adding to his/her decision making capacity. So information visualization in a way, shifts the focus of information retrieval to information processing from the lexical to the spatial and visual sphere.

Visual search then - used either for web information retrieval, or for non-internet information retrieval, is the ability to browse search results by using 2D or 3D color graphics and animation. These search results can reveal the structure of information giving it a spatial dimension - allowing users to navigate and interact with it a completely different way than text based results.

It is important to note here that we are not talking about image or multimedia retrieval/search, we are talking about retrieving textual information by using visual means.

Visual Search on the Internet: How Do We evaluate?

Against the backdrop of the definition of visual search, let us lay down a few evaluation criteria to judge visual search on the Internet:

• The basis for relationship grouping of information results represented by images
• Graphing/layering these visually represented information clusters to maximize the attention span of the user
• Cuing users to maximize the navigation capabilities of the visual presentation of information (text labels, notes, etc.)
• The type of visualization used
• Being able to navigate from visual search to text base search
• The ability to derive actionable information intelligence

Here are a few examples of visual search on the Internet:

• KarToo
• Anacubis
• WebBrain

Anacubis, Kartoo and WebBrain: A Comparative Chart

1. Evaluation Criterion : The basis for relationship grouping of information results represented by images

• Anacubis:Keyword based grouping of Google Results
• KarToo: Uses an amalgamation of keyword based and syntax based techniques to group search results
• WebBrain: Text Images based on Open Directory projects. Uses the Open Directory relationships to show the visual text map.

2. Evaluation Criterion : Graphing/layering these visually represented information clusters to maximize the attention span of the user

• Anacubis: Very difficult to see the images and the text not a very intuitive way to visualize information- at least in the Google demo.
• KarToo:Interesting animations, and also interesting attempt at relationship mapping, but too jazzy to aid the human attention span
• WebBrain:Very clear structure but not much of graphing.

3. Evaluation Criterion : Cuing users to maximize the navigation capabilities of the visual presentation of information (text labels, notes, etc.)

• Anacubis:Title of the web page and the URL
• KarToo:Top keywords are shown in the right hand corner on the mouse-over
• WebBrain:No cues, the images are themselves text-based

4. Evaluation Criterion : The type of visualization used

• Anacubis: Relationship Graph
• KarToo: Flash Animation relationship graph
• WebBrain:Relationship Graph

5. Evaluation Criterion : Being able to navigate from visual search to text base search

• Anacubis: Click
• KarToo: Click
• WebBrain:Results are shown in a different frame.

6. Evaluation Criterion : The ability derive actionable information intelligence

• Anacubis: Almost none in the Web-based demo
• KarToo: Almost none
• WebBrain: Almost none

Besides the above mentioned players, there are few more interesting ideas floating around that come very close to the metaphor of visual search:

• Google Browser: This browser uses Google Similar Pages or Related feature to graphically show related links to a webpage. The company TouchGraph also has other interesting demos that show related searches on Amazon and a relationship map showing linkages between LiveLink Journals.
• Browse3D: Browse3D is a different type of web browser that provides a graphical three-walled interface for loading and using multiple websites at the same time. Broswer3D in a way is a take on the concept of information visualization.




• Google Viewer: Another take on information visualization concept, but this time this Google Viewer aids another cogitative faculty: ability to fast-read data.




• MapStan: Another example of a technology using Google Related to visualize searches in the form of information cartography. "MapStan designs and integrates information cartography tools and technologies based on the analysis of structured information (habits, behaviors, etc.), synthesizing a large mass of information into a set of interactive visual representations"

Information Visualization: Other Players

Apart from the demos that we have seen above, there are other players who are providing visual search and more advanced information management and retrieval solutions to businesses. Notable amongst them Antarctica Systems and Inxight . Antarctica Systems had a number of demos on line but most of them have been pulled back, most probably for competitive intelligence reasons. Still a few remain online Map Net and Canada-VC Deal Map are good examples representing Antarctica technology.

Inxight on the other hand, is one the most respected names in unstructured information management field and its StarTree supported by the VizServer platform is one of the most innovative solutions in this arena. It is important to note that StarTree (click here to see a few demos of StarTree) comes from the Hyperbolic Browser developed by Xerox Parc Research.

In terms of the evaluation criteria, both the solutions are way ahead of the visual search examples mentioned earlier- Antarctica Systems' Visual Net and Inxight's Star Tree solutions provide a technologically advanced platform that can help users in identifying, discovering and providing a visually conducive environment to act upon the actionable information intelligence derived from visual information clustering.