Visualizing Search Results

After a search query is entered into an information retrieval system, the system responds with a report of the retrieved items. Traditional database systems typically present a list of items that can be sorted by certain elements. On the Web, the display generally consists of an unorganized list, ranked according to what the search engine determined to be relevant. The information given for each entry is meant to give the searcher an idea of the item’s content so that he/she can decide whether or not to obtain it. While this may be adequate for some searches, there are several weaknesses to the list approach:

• It is not informative for the user because it gives equal weight to all results (Miller, 2004).
• Handling large results sets is cumbersome (Fagan, 2006).
• Few end users ever go beyond the first few page of results, even when there may be thousands of links related to their query (Jacsó, 2005).

Is a picture worth a thousand words?
Data visualization tools can provide an overall view of the retrieved items and how they relate to the topic in question, something that simple lists cannot do. Luther, Kelly, and Beagle (2005) noted that tools that combined visualization with clustering helped users discover aspects of the topic they did not know, which in turn led to more meaningful searches. Going back to our needles and haystacks, visualizations address the problem of the "unknowns" by helping us view relationships among data, without having to know ahead of time that those relationships exist.

Several studies have tested the efficacy of visual search tools with mixed results. In one study, users found the visual system interesting, but didn’t trust it (Fagan). Visual search engines do have a lot of pizzazz, but they also need to demonstrate sound, reliable functioning in order to win the trust of users.

I rather like visual search tools, but that might be simply because I am a visually oriented person. These are the visual search tool I find most intriguing. If you are not familiar with some of these, give them a try. Or leave a comment with your own favorites.

AquaBrowser divides the search results into three panes. A traditional list is in the center. On the left is presented avisual map of related terms and homonyms. On the right are category options to filter the results.

KartOO is a metasearch engine that displays the results as a map or cloud where each result is represented by its URL and a thumbnail of the site. The cloud also contains keywords that show how the items are related.

KoolTorch clusters search results and applies a taxonomy along with the graphic display.

Grokker – my personal favorite - displays results as a series of categories set in a circular map. One of the best features of Grokker is that the categories can be edited, changed, or renamed to personalize your map.

Maramushi Newsmap is a visualization of the Google News aggregator. Is not a search engine, but it is a type of information retrieval system. News stories are presented in a grid to indicate the amount of content being generated. Headlines are displayed by category, volume (the more articles there are, the bigger the block), and currency (more recent items are brighter.) I have to admit that when I first saw Newsmap, I thought it was preposterous, but now I find it rather addictive.

References

Fagan, J.C. (2006). Usability testing of a large, multidisciplinary library database: Basic search and visual search. Information Technology and Libraries, 25(3), 140-150. Retrieved April 8, 2008, from Wilson OmniFile FT Mega Edition database.

Jacsó, P. (2005). Relevance in the eye of the search software. Online Information Review, 29(6), 676-682. Retrieved April 8, 2008, from Emerald database.

Luther, J., Kelly. M., & Beagle, D. (2005). Visualize this. Library Journal, 30(4), 34-37. Retrieved April 9, 2008, from EBSCOhost Academic Search Premiere

Miller, R. (2004). Get the picture. EContent, 27(4), 30-35. Retrieved April 9, 2008, from Gale Expanded Academic ASAP.