Anselin & Getis give a good overview of the issues that pertain to the integration of Spatial analysis into GIS. Although it is quite a dated paper(1992), it does a good job of highlighting exactly why better integration wold be beneficial for the entire field of GIS. As noted by the authors, one of the key functions of a GIS is the analysis portion, which in turn encompasses spatial statistical analysis. They correctly identify this function as vital for more complex and in depth case studies in the future.
Technology has evolved a great deal since the time Anselin & Getis wrote their review. Modern day GIS now include many spatial statistic tools built right into the system. For example, for several courses I have used the “spatial analyst” toolbox in ArcGIS to perform statistical analysis of raster datasets. This toolbox holds a wide array of functions, ranging from calculating the statistics of objects in a raster (zonal statistics), to combining different rasters based on the measure of central tendency of the data (cell statistics). In addition, there are now even excellent standalone programs made to specifically analyze statistics. Some of these programs, such as R, allow the user to perform complex statistical analysis of datasets. In addition, more complex programs, such as STATA, include a spatial component that allows the user to perform spatial statistical analysis on datasets.
Overall, the authors do a good job of providing an overview of the problems, as well as the benefits of better integration between spatial statistical analysis and GIS. However, many of the issues raised throughout their paper have been solved over the years, with the evolution and growing complexity of GIS. Spatial statistical analysis is an important component in any GIS. In the present day we have the ability the perform complex spatial analysis within GIS programs such as ArcGIS, something Anselin & Getis could only dream about at their time of writing their paper.
-Victor Manuel
Tags: GEOG 506