Database visualization differs from other types of information visualization due to the diverse nature of the data stored in the database. The attributes used to organize the presentation may not have a regular scale and, in fact, may not even have a semantically meaningful order. Data can be categorized based on whether it has inherent order or scale. For example, numeric data has both an order and a scale. In contrast, geographic data (latitude, longitude) can be easily scaled according to a well-defined metric, yet ordering is not as straightforward. This problem is further complicated by the frequent use of reference values in database systems, for which the order and scale are defined by an external application and cannot be inferred by the type of data.
This thesis presents a system architecture to address the problem of database visualization. The essential and unique component of this architecture is the mapping functionality, which adds order and/or scale to the data. A map may take into consideration the domain of the data as well as other information. The features of a particular map can be programmed to support a wide variety of visualizations. Naturally, the architecture supports the addition of new maps in a modular fashion. In addition, the architecture contains the following more conventional components: query specification, database storage, filtering, plotting and image display.
In addition to the development of the architecture, we propose to analyze the usability of the database visualization system by conducting an in-depth usability evaluation. The evaluation will be structured to measure the effectiveness and efficiency of users solving realistic problems using the mapping language.