Facet-based Information Navigation

Facet-based site navigation systems provide a way of browsing complex data sets without forcing a formal taxonomy onto the data. Facet-Based browsing is different than other methods of crawling information. A common system for navigation is a Taxonomy where you start at the top of the tree and browse your way down to the information that you are looking for. Many different web sites use a taxonomy for browsing information. At ConsumerReports.org, you click on Appliances -> Large Kitchen -> Freezers. Another example would be looking for a MedLine article in a physical library. You find the Journal -> Volume -> Issue -> Pages.

In a facet-based system, you could select one of several facets to crawl the information by: Date, Author, Journal, Keyword. You can click and limit the results by each Facet. Browsing a Facetted system will never lead you to an empty result. You can see a demo at Siderian Software’s Seamark Demo. Siderean Software’s products crawl the repository to create a facet-based information browser. The selections are generated on-the-fly with each sub-selection.

Endeca also has a facet-based browsing system. They built a demo that dynamically crawls and displays 90,000 wine reviews from Wine Spectator. You can crawl (sort) by any of nine different facets: wine type, winery, country, price and year.

Tags and browsing by tags (like at del.icio.us) allows for a very loose method of coalescing information. The social aspects of tagged based systems is critical (I think) to building an Enterprise Collaboration and Communication environment. The problems of tag-based systems are that: you are unsure if you found everything related to a tag, there will be inconsistency in the application of tags, stemming (various tags that mean the same thing like Mac, MacIntosh) causes missed opportunities and tags with multiple meanings cannot be teased apart without context.

Formal taxonomies are difficult for users to navigate unless they are fully educated in the taxonomy. They are difficult to apply to information especially in the very loose conversations that make up collaboration and communication.

Facet-based browsers could add the structure that allows for grouping similar information and filtering large data sets without constricting the flow of information. One very interesting demo is Facetious – a tag based browser for del.icio.us. The facets here are: Tag, Creator, Site, Feed and Date.

Facet-Based browsing could fill the middle ground between the folksonomy based on tagging and the formal taxonomy in Enterprise Information Management.