Almost all GIS packages provide a range of facilities for computing simple univariate statistical measures on tabular attribute data associated with vector objects. Many of the basic statistics listed in Table 1‑3 are provided, either directly by opening an attribute table and selecting a column for analysis, or by executing some form of simple calculation facility. In addition to these basic measures, tools are often provided that display frequency histograms of data, with or without data transformation. Such facilities may be stand-alone or may be embedded in related functionality, such as map classification and symbology tools.
For image or grid datasets many packages provide a wide range of non-spatial and spatial statistical facilities. Purely non-spatial facilities include the same type of statistical measures as for vector objects, where the attributes are grid cell values rather than attribute table columns. In addition, many packages provide a range of statistical tools designed for multiple grid analysis. A number of these go beyond data description to provide facilities such as data reduction (examining data redundancy) and modeling (e.g. simple regression techniques). Some of these are described in the following subsections.
Much of the subject commonly described as “Spatial Statistics” deals with vector datasets, and many of the tools and techniques that have been developed apply (directly or indirectly) to point rather than line or area-based data. Increasingly, however, integrated tools such as GeoDa, PySal and SAM have become available that facilitate the direct exploration, analysis and modeling of zone-based data and small area (local) patterns.