Selection

A vital task in creating effective graphics for visually relating ideas is the filtering of vast fields of data into a units that are simple enough to be understood quickly and intuitively but detailed enough to convey useful information. Often, creators of visual displays are too ambitious with the amount or detail of data that they attempt to present on screen. Make sure that you select and generalize if necessary.

Recall that, in all visualization applications, one of the primary considerations is the audience for whom the display is being created. To this end, a designer should carefully select those themes or variables that are most appropriate for the communication of ideas. If, as in many cases, the audience a single GIS specialist who is exploring data privately, searching for connections, trends, and patterns, it is likely that the display of highly detailed information is most important. Even in this case, however, proper visualization techniques as outlined in this unit will make data exploration faster and more accurate. However, for public presentation, only those data that are immediately relevant and quickly understood should be included.

An example of this type of selection is the display of the results of a study of potential development sites for a new housing complex. A group of GIS specialists may have the street and hydrology network, with labels, an elevation model, and other layers of information on hand to make a decision. When exploring this data, they might create a complex multivariate map with symbolization techniques known to them alone. However, when presenting this data to a planning commission or contractor, they should limit their display to a small number of variables -- like surrounding streets, streams, parks, and schools.

Generalization

Such refinement may be taken a step further by generalizing the data both visually and mathematically. Visual generalization might involve, for example, including only those streets that give a viewer reference to the location of the housing complex, or perhaps labeling only the major streets in the vicinity. Other visual generalization techniques involve aggregating objects as scale is decreased (such as transforming 10,000 buildings into an "urban" area then into a dot as the geographical region of the display is increased), or including only those features with an attribute greater than some threshold (such as displaying only those airports with runways over 3000 feet).

Mathematical generalization techniques vary from simple to extraordinarily complex. In most GIS applications, linework generalization is the most important application of these techniques. It is seldom necessary in practical situations to carry through the most detailed linework of coastlines, rivers, or highly irregular political boundaries. As mentioned, such detail may be not only unnecessary but also distracting. In vector GIS, this linework is stored and displayed as an exhaustive series of line segments connected at points. Selectively eliminating, for example, three of every four of these points creates a smaller database and a less complicated display while still maintaining enough visual information for recognition and understanding.