Large hierarchies, such as trees or variants thereof, require complex interaction models as they typically span beyond the available display space. Space-filling visualizations, such as the TreeMap, are well-suited for displaying large hierarchies in limited viewing space. They are also designed to display the properties of nodes in hierarchies in space-filling visualizations. To browse the contents of the hierarchy, the primary mode of interaction is by drilling-down through many successive layers. In this thesis I introduce a distortion algorithm based on fisheye and continuous zooming techniques for browsing and searching data in space-filling representations, such as the TreeMap. The motivation behind the distortion approach is for assisting users to rapidly browse information displayed in the TreeMap without opening successive layers of the hierarchy. For searching tasks the distortion technique assists users in identifying the search results even when the hierarchy is dense and is capable of conveying importance level of search results.

Three experiments were conducted to evaluate the new approach. In the first experiment (N=20) the distortion approach is compared to the drill-down method. Results show that subjects are quicker and more accurate in locating targets of interest using the distortion method. The second experiment (N=12) evaluates the effectiveness of the distortion technique in a task requiring context, we define as the context browsing task. The results show that subjects are quicker and more accurate in locating targets with the distortion technique in the context browsing task. The results of both these experiments provide strong evidence that distortion based techniques applied to space-filling visualization facilitates rapid browsing. The last experiment (N=12) evaluates the effectiveness of the distortion technique for presenting search results. The results do not show any improvement in the distortion method over currently available techniques for presenting search results.