2005 02 23
GIS: Modeling a Better World
MapQuest and OnStar are two example applications of GIS -- Geospatial Information System, a term that effectively explains itself. Most business information is stored, one record at a time, in tables, and may be presented in (web) documents, tables, and charts. A more visual way of managing information links data in tables to a model of Earth.
If a data record has field values for latitude and longitude, the record can be linked with a graphic object depicting a feature on a model of Earth--appearing as a 2-dimensional map, or even 3-dimensional planet. The interface to the model includes the expected navigational tools (panning and zooming). Another type of location data is a street address. The address string is parsed into components and located by matching to a record in a street segment table that already has been located. This is called geocoding - putting data on a map.
Data, 2-way linked with a spatial representation, is the basis of a Geospatial Information System. Create a new record in a table by simply drawing a new (vector) feature on the map. The character data "lurks beneath" the map, which is the interface enabling ad hoc revelation of the "hidden" information. Click a map feature and up pops info from the linked record located on the map. How features are depicted depends on user choices of symbology such as color, shape, and size. A thematic map results when the data "drives" the symbology, based on values in one or more fields of the table.
For example, to represent census data geographically, first chose the field value for Income to determine the color of each census tract (record) polygon. Then choose a color range: light yellow to dark green. The thematic map portrays high-income tracts as dark green, contrasted with lighter and yellow, lower-income tracts.
Each feature layer on the map corresponds to a table. The number of possible layers is unlimited. While the visibility of each layer may be clicked off and on, the data “behind” the layer is always available to support a variety of display and analysis functions. For example, by adding a layer of zip code polygons we may select only the high-income zip codes for a mail campaign.
Today, the organization (government, utility, business, …) is building its GIS to represent all features that are of interest, layer by layer, including, e.g.: roads, surface water, parcels, sewers and buildings. Generally a background photo-image completes the GIS. Unlike feature layers, an image layer is not linked to a table, but may be upgraded to a grid surface, a sort of image spreadsheet. Grid layers support “geomath” between surfaces producing a new analytic result layer.
Feature tables may have as many data fields as desired. These "smart features" fuel the power of the GIS user, selecting only features that satisfy criteria (census tract records where Income is greater than 50,000); then applying that layer to (selected) features in another layer, such as zip code. GIS facilitates multi-step, multi-path (algorithmic) geoprocessing to determine, for example, based on available data, where to put the next store. (in high-income tract, close to interstate, zoned Commercial, ...)
Like all software technology, GIS evolves from desktop app to web service, finding synergies with evolving internet standards based on XML like BPEL (Business Process Extension Language). As the internet-facilitated language of GIS reaches the myriad avenues of human concern, a more holistic understanding of world is enabled in every specialized endeavor. Leading the GIS march since 1970, Environmental Systems Research Institute (ESRI), based outside of Los Angeles, provides ArcGIS, the most widely used GIS application in the tri-state area.
john r. schmidt, M.S., president of NCAD Corporation, and ESRI Business Partner, since 1989.