Abstract
A major application of computer mapping is to display levels of some dependent variable over a geographical area. In general, this dependent variable is known only at a limited number of specified locations, henceforth termed “data points,” and must be interpolated for other locations. The need for an interpolation algorithm arises in a great variety of applications. In city, regional and geographical planning, social and economic characteristics (e.g., population density) are known only at the centers of data collection zones (e.g., towns or census tracts). In meteorology and air pollution analysis, atmospheric conditions and pollution concentration are known only at monitoring stations. In geology and oceanography, depths are known only where soundings have been made. In epidemiology, the death rates are known only for some representative point (or as an overall average) for the state or county over which vital statistics have been calculated. The values of the dependent variable to be mapped (subsequently denoted by Z) form a surface over a geographical area in the same way that a relief map represents a topographic surface.