Augmenting search with a Semantic Visual Graph

Improving search in a large database by using a SVG visualization of a reduced Xml Topic Map

Edwin de Jonge
R&D Informatics
Statistics Netherlands
Prinses Beatrixlaan 428
ejne@cbs.nl

Abstract

Statistics Netherlands produces statistics on a wide spectrum of topics about Dutch society. All statistical results are published for free on the internet via the StatLine database, currently containing over 100 million statistical facts. Searching and finding information in such a large database is not trivial. This searching problem is aggravated because statitical institutes traditionally treat the statistical areas of interest differently. For example economical and social statistics are rarely made by the same statisticians. Many statistical areas however overlap and relate to one other. Each statistical overlapping area though has its own jargon, homonyms and synonyms, which makes searching difficult.

These problems can be alleviated using semantic webtechnology (XTM) and SVG visualization. This paper describes briefly what concept of semantic searching we use and then focuses on SVG visualization and implementation: our search is augmented with a Semantic Visual Graph (SVG). Scalable Vector Graphics is a the natural choice for its implementation because we need animation, textPath and callback functionality. The paper shows with an example that visualization is an integral part of our semantisch searching process. The visualization allows a user not only to specify his search but also to search associatively, allowing the user to find information more intuitively.