As we have already seen in issue number 26 the semantic web is an initiative led by Tim Berners-Lee to grant access to the information residing in the Net not only to humans but also to machines.

Some companies are beginning to use the semantic contents (the meaning of their contents) that has begun to be codified in the web sites in order to improve the search procedures of relevant information. Some examples of this are Amazon.com or Yahoo shopping. But the semantic contents of a web site can be used advantageously for the visualisation of its information, since we already have the knowledge about the meaning and the existing relations between the concepts used in it.

Among the companies that are working in this direction in Europe we find Aidministrator, a Dutch company devoted to information presentation & distribution that has developed an interesting visualisation system based on the semantic contents (the ontology) of a web site. 

The approach that Aidministrator takes is to consider that in any visualisation procedure you begin with the raw data. This data is then organised, filtered and classified until it constitutes a semantic structure that provides a coherent meaning. Once this is accomplished the obtained structure serves as the basis to produce the visualisation. 

Fig. 1: In any visualisation process you begin with the raw data. These data is tabulated, filtered and classified until a coherent semantic structure is obtained. This allows you to viusalise the structure using a convenient visual metaphor.  (Sketch based in the article cited at the end of the page) Click on the image to see it enlarged.

As they themselves say in their paper "Ontology based Information Visualisation"* this is a "generic procedure that is commonly used in information presentation tasks, ranging from constructing a lecture to designing a website", first you decide what to show (the structure) and then how to do it (the visualisation).

The Aidministrator system represents the ontology codified in a web site, (see the attached  images), as a system of text labels and spheres connected by straight lines. 

• Each label represents a class in the ontology (a category of pages, for example, sports). 
  
  
• Each sphere is an instance of a class (every page belonging to that category). 
  
  
• The lines indicate that an instance is a member of a class or that a class is a subclass of another one (for example football is a member of sports and sports could be a subclass of entertainment).

All these elements are located in space through a system of attraction and repulsion between the objects (as if you had springs inside the lines) This system keeps them in a dynamic balance that makes the objects that are semantically close appear spatially close, and the objects that are semantically far away appear in distant locations.

In this context "semantically close" means that two classes share many instances or that two instances belong to the same class.

Figure 2: The visualisation of the web of a job-agency. Click on the image to see it enlarged. Image courtesy of Aidministrator	Figure 3: The application of a new version of the algorithm to a database of documents on educational issues. Click on the image to see it enlarged. Image courtesy of Aidministrator

For example, Fig. 2 depicts the contents of a Dutch web site of a job-agency. The ontology classifies the jobs by economic sector: recreation, finance, education, etc. Visualising this ontology using the visual metaphor of the spheres connected by "springs" allows us to immediately detect some trends and patterns as well as outliers.

• The size and intersections of the classes stands out immediately due to the clustering of spheres. 
  
  
• You can also see that quite a few of the jobs are classified under several classes; for example one of them "hangs" from finance, management and secretarial. This gives us a good idea as to whether it is relevant to us or not. 
  
  
• Finally we can see that classes placed on opposite sides of the diagram don't have any member in common (government & security vs., healthcare & sports).

The objective of Aidminstrator is to use this visualisation technique for:

•  supporting navigation on websites or other document collections, possibly using multiple views on the same data
  
  
• providing overviews of data sets for analysis tasks, possibly over time (see how a map changes over time)
  
  
• supporting querying (on search engines or regular databases) by visualising the structure of a result set in a more clever way than just providing a list or table.

The semantic web is beginning to be deployed timidly. As in other fields, the knowledge of the semantic contents will help enormously to visualise the information. As we have seen there are already companies working on that.

* Ontology based Information Visualisation, Proceedings of the Vth International Conference on Information Visualisation, London July 25-27, 2001, IEEE 

Links of this issue: