Abstract:
|
Paradigms and methods for visually exploring complex datasets have been simultaneously developed over the last several years by many different research communities, including statistics, cartography, computer science, and engineering. These communities have their own particular purpose, problems, and biases. Some of these techniques emphasize the correlation among data dimensions, others are designed to find clusters or outliers and still others attempt to provide holistic views onto highly multivariate data. For example, the recent interest in data mining and knowledge discovery has led to new visualization tools to support analysis of massive datasets in the search for association and clustering. Visualization paradigms and methods rely on a variety of different perceptual and cognitive principles in order to be effective; that is to say, they take different approaches to the problem of engaging the perceptual and cognitive systems of humans. Effective visualization demands careful consideration of the perceptual and cognitive demands placed on the user, and the task the user is trying to accomplish.
|