Spatial Statistics in SAS

Abstract

From the abacus to the adding machine to the supercomputer, for centuries humans have used aids to enable mathematical computations. As the mathematical tabulations grew in complexity, so did the ‘machines’ that enabled more complex calculations. This in turn presented the problem of implementing beautifully written formulas in a form a computer ‘aid’ could understand. Today statistics specifically has a huge variety of software implementations available to choose from, some of which focus on a specific subdiscipline of statistics, while others encompass statistics more broadly. SAS Institute, as did many specialized software companies, evolved from an academic background in partnership with IBM, and its statistical package is used widely in statistics as well as a plethora of disciplines that rely on statistical results. Here we describe some of the ways SAS has been used in the past for spatial statistics, and some of the more recent additions made to explicitly include spatial information and geographic visualization, and give two SAS implementation examples, the calculation of Moran’s I and the eigenvector spatial filtering spatial statistical technique.