There are really two parts to ArcGIS geoprocessing: the framework and the tools. ArcGIS has a fully developed framework for performing geographic processing. The fundamental model is based on the simple idea of data transformation.
Data + Tool = Data. Parameters can be used to control the interaction between the data and tools. Data and tools can be strung together to create quite sophisticated models. Here is a relatively simple model for calculating hot spots (clusters) in crime events.
Models can be created using visual programming (drag ‘n drop data and tools on to a canvas, connect them together, specify parameters and run) and scripting (Python, VB, etc.). The geoprocessing framework is available ubiquitously throughout ArcGIS (on the desktop, as server web services, and embedded in a desktop run-time engine).
There are hundreds of tools in the ArcGIS 9.2 release that are grouped into toolboxes. Many of the toolboxes are in the core, and some are added when extensions (3D, Spatial, etc.) are licensed. Any user can develop their own tool (by combining existing tools, or adding new code).
This apparently simple approach to geographic processing belies a very powerful and sophisticated capability for spatial analysis and modeling. When combined with the power of ArcGIS is offers some very interesting options for analysts and modelers.
Data management for very large, multi-user data sets. In the past, spatial analysis and modeling projects have used relatively small data sets (thousand of features, grids with less than 100,000 pixels) because of the difficulty of managing data. ArcGIS’s geodatabase takes care of this and can store all data in a DBMS or file system.
Tools for data conversion, transformation and integration. ArcGIS 9.2 offers a full spectrum of basic utility tools to convert, transform and integrate data (change projection, convert spatial base from, say, polygon to point centroid, edge match separate tiles, etc.) which takes a lot of the hard labor out of geoprocessing.
Mapping and visualization environment. The ArcMap and ArcGlobe 2D/3D mapping and visualization environments support data exploration, scientific visualization and communication of final outputs. New 9.2 scientific charting and graphing functions also help with data understanding and visualization. ArcGIS allows GIS to be inserted into policy frameworks so that it can be effectively consumed.
Scripting and customization. The geoprocessing framework provides support for scripting models using both Python and .Net / Java languages. This gives programmers precise control over model execution and allows for easy extensibility.
Integration with external systems. Recognizing that it is not sensible to model everything, every time inside a GIS, the geoprocessing framework has lots of options for close- and loose-coupling with external systems such as SAS, SPSS, R, and any data source or callable program.
In summary, ArcGIS offers a simple, but very sophisticated geoprocessing framework and set of tools. Prior to 9.2 these tools were really only able to be applied to static map style spatial analysis and modeling. The next post will look how simulation and advanced modeling can be performed with ArcGIS 9.2.