General Concepts

Analysis is always focused on a certain subject, which is a thing or phenomenon that needs to be understood and, possibly, modelled. The data science process involves analysis of three different subjects: data, real world phenomena portrayed in the data, and computer models derived from the data. A subject can be seen as a system composed of multiple components linked by relationships. Understanding of a subject requires understanding of the relationships between its components. Relationships are reflected in data, which specify correspondences between elements of different components. Relationships are studied by identifying patterns in distributions of elements of some components over elements of other components. A pattern consists of multiple correspondences between elements that are perceived or represented in an integrated way as a single object. In visual displays, patterns appear as shapes or arrangements composed of visual marks. Depending on the kind of components (discrete entities, time, space, numeric measures, or qualities), different types of distributions can be considered: frequency distribution, temporal and spatial distributions, and joint distribution of several components. Each type of distribution may have its specific types of patterns. We give special attention to spatial distribution, noting that the distribution of visual marks positioned in a display space is perceived as a spatial distribution, in which a human observer can intuitively identify spatial patterns. This inclination of human perception to seeing spatial patterns can be exploited by creating artificial spaces where distances represent the strengths of some non-spatial relationships.