Abstract
Recent years have seen the emergence of the World Wide Web as a significant resource that is increasingly relied upon for both the dissemination and gathering of information. However, information on a specific topic is often distributed across independent web pages stored at different geographical locations that are updated and restructured on an ongoing basis. Therefore, users with an interest in a particular topic may find it difficult to locate information, monitor changes in the web pages and make coherent observations about the way that the information is evolving over time.In this paper we propose a framework to support users in gathering information from given web pages based upon topics of interest. This is achieved by creating abstractions of user-defined web spaces that are based upon both content and structure. Importantly, the abstractions are maintained over time in a format that supports the analysis of how the web space is evolving. The abstraction approach uses clustering to model the content of pages whilst the structure of links between pages is recorded in a database. Visual interfaces can then be employed to allow users to easily comprehend the web space.