Dynamic queries is an approach to database querying, information exploration, and information retrieval. Building on principles of direct manipulation, dynamic queries utilizes visualizations and graphical query devices to allow users to visually explore, browse, and query large datasets. This dissertation introduces dynamic queries and explores a whole range of research issues related to it.
When dealing with large sets of information, tasks include locating pieces of information in huge databases of objects, numbers and documents, exploring trends and patterns in large data sets, identifying anomalies and exceptions, discovering clusters of information over time and other dimensions, etc. These tasks are difficult to perform, not only because of the sheer size of the data, but also in that the tasks are exploratory in nature with ill-formulated goals which change over time as users encounter new information and new views of old information. Dynamic queries tries to battle these problems through the introduction of a combination of visualizations, graphical query devices, details-on-demand, and tight coupling.
Dynamic queries is initially evaluated in a controlled experiment where 18 subjects interacted with three different interfaces, including one based on dynamic queries, and dynamic queries was found to the most effective - both in terms of time to solve tasks and the number of errors. Dynamic queries builds heavily on so called query devices which are graphical interaction objects used for formulating database queries, such as rangesliders, alphasliders, and toggles. The design space of query devices is explored in a semi-formal way. One of these query devices, the alphaslider, is elaborated upon and several designs are compared in a controlled experiment with 24 subjects. An important concept for direct manipulation based exploration systems is tight coupling which helps users avoiding empty query results and provides navigation support in the space of possible database queries. Finally, the Information Visualization & Exploration System (IVEE) is introduced. IVEE can create dynamic queries environments automatically from practically any dataset given as a relation and explores the use of dynamic queries together with many other interesting techniques in an end user system.