GVIS: A framework for graphical mashups of heterogeneous sources to support data interpretation

This paper introduces the GVIS framework and describes one of its applications built in support of user profile awareness. This application is aimed at opening part of users' profiles to their inspection by exploiting a graphical representation of their personal data. We developed an infrastructure for presenting these high level information in a configurable and adaptable way. The framework we developed is able to retrieve data from heterogeneous sources just by writing a small adapter and allows us to mix together different streams through an XML configuration that relies on a set of operations for elicitation of the most interesting fragments. The final goal is to provide an easily readable graphical representation of the most relevant information, in order to support the human visual system, more capable to have an overview with this kind of solution than with text. As an example application we have mashed up URLs from user browsing history with tags coming from del.icio.us: the resulting output, represented as a pie chart, shows the most relevant subjects followed by a user. Some open issues and problems, we hope to research next, are presented in the conclusion part.