Most of the data in BI systems is multidimensional in nature. How do we present multidimensional data in a way that is clear, meaningful, and easy to understand? The design of visualization involves the effective use of fundamental visual elements and visual properties.
Visual elements are the basic building blocks in a chart or diagram to visualize data items. The most fundamental and abstract elements are: point, line, surface (area), and volume (3D). These basic elements, and the more complex elements built up on them, can represent almost anything in a visualization.
Visual properties or variables are used to "decorate" visual elements, so that the values or meanings of data items can be directly and easily perceived and understood by human. The most commonly used properties can be summarized as SCOPeS: Size, Color, Orientation, (spatial) Position, Texture, and Shape:
- Size: the size of an element is an important property used for continuous data values. It can be implemented as length, width, height, area, angle, etc. For various reasons, it is common that the size property does not directly and truly represent the underlying value. In these cases, it must be very careful to design the size property, because unreasonable distortions will impact human perception.
- Color: color is the most common visual property used for both categorical data and continuous data. It also include hue, brightness, and gray scale.
- Orientation: it is closely related to shapes, and can be seen as variations of a particular shape or pattern. An common example is arrows or hands pointing to different directions.
- Position: data values can be visualized as absolute positions in the visualization, or as the relative distance between elements. It is commonly used to visualize the position of data items against a pre-established scheme (such as a Cartesian coordinate system), categorization and grouping of date items in terms of similarities and differences, or spatial distances (especially used with maps).
- Texture: texture is important when color sensitivity is an issue. Implementations include fill patterns, border patterns, shadow, etc.
- Shape: shape can be applied to any visual elements. When used to visualize individual objects or data items, it usually represents nominal or categorical data values.