GEOGRAPHIC INFORMATION SYSTEMS Michael F. Goodchild A geographic information system is designed to capture, store, display, communicate, transform, analyze, and archive georeferenced information, that is, information tied to specific locations on the Earth’s surface. Geographic information systems enhance and to some extent replace the traditional role played by maps, but are also capable of handling information in the form of satellite images of the Earth’s surface, as well as information from surveys and administrative records that have been georeferenced. They are increasingly used in the social sciences to support research based on cross-sectional data, or studies for which geographic location and context are important and useful. I. Introduction The origins of geographic information systems can be traced to the mid 1960s. Early computers were designed primarily for numerical processing, following the lead of Babbage and others. But by 1965 other applications had begun to appear, supported in part by the development of specialized peripherals, notably pen plotters and map digitizers. About this time the Canada Land Inventory faced a massive problem of map data processing: how to take the very large number of maps created to document Canada’s underutilized land resource, and to produce tables of the amounts of land available for various types of development and use. Measurement of area from maps had always been time consuming, tedious, and inaccurate when performed by hand. But if the maps could be converted to digital form, simple algorithms would allow areas to be measured and tabulated electronically. The Canada Geographic Information System (CGIS) was thus a response to a well-defined need.By the 1980s, commercial GISs had begun to appear, offering a wide range of functions that in various ways were too complex, tedious, inaccurate, or expensive for humans to perform by hand. These included simple measurement of area and length, transformations needed to alter data formats, simple statistical analyses such as the calculation of means and standard deviations, and a host of more complex and sophisticated methods generally termed spatial analysis. In addition, GISs were provided with advanced capabilities for data display, including mapping and various forms of data visualization. The scientific community was quick to recognize the potential of GIS, and through the 1980s and 1990s GIS emerged as an indispensable tool for research in any discipline dealing with the surface of the Earth, or the near-surface. In the social sciences some of the first applications were in archaeology (*), but political scientists, criminologists, demographers, and epidemiologists have also been prominent early adopters. For extensive bibliographies covering applications of GIS and spatial analysis in the social sciences, as well as information on sources of tools, data, and other resources see the web site of the Center for Spatially Integrated Social Science (www.csiss.org). In recent years GIS has undergone significant transformation, as applications have emerged that go well beyond the early notion of a digital assistant performing tasks that humans find difficult. The advent of the Internet and the WWW had by 1995 induced a sharp change of perspective, in which GIS was viewed as a means for sharing information between people, in addition to its more traditional role. Many web sites were created offering to supply visitors with geographic data sets, or to create maps on demand, or to perform simple GIS services to user specifications, using data provided by the user or by the site. Maps are compelling ways of presenting information, and spatialanalysis has been reinterpreted in recent years as a set of methods by which one person adds value to information, by making visible what might otherwise be invisible to another person, thus strengthening the message. For a comprehensive survey of Internet GIS see the introductory GIS text by Longley et al. (2001). Most recently, advances in technology have brought the promise of GIS that is no longer confined to the office, but carried into the field in the form of portable and wearable devices. Wireless communication is available to download and upload data to and from Internet sites, and sufficient power is now available in portable devices to support virtually any GIS operation. The advent of field GIS offers to revolutionize the nature and practice of field work, in social surveys and other field-based social science. II. Representation At the heart of a GIS is a system of representation, by which features in the real world are coded in the binary alphabet of the digital computer. GIS representations typically include three aspects of real-world features: their locations on the Earth’s surface, using a convenient coordinate system such as latitude and longitude; their attributes, or the things that are known about them; and any relationships of importance between them. Examples of relationships include adjacency, such as the adjacency that might exist between two neighborhoods; and connectivity, such as the connections that might exist between parts of a street network. Attributes provide much of the richness of a GIS representation, especially in the social sciences. Reporting zones such as census tracts might carry large numbers of descriptive attributes, created from the summary tables provided by the Census, such as average income or percent unemployed. Individual points representing the locations ofindividuals in a sample survey might carry as attributes the information collected in the survey from each individual. Underlying the representation of geographic variation are two distinct conceptualizations. In the first, the features on the Earth’s surface are discrete objects, much as a tabletop might be littered by books, pens, or coffee mugs. Discrete objects can overlap, and empty space can exist between them. This discrete object view is particularly appropriate in the representation of moving or persistent objects, such as individual people or vehicles. Objects can be represented as points, lines, or areas depending on their size in relation to the geographic extent of the representation. Areas are represented as sequences of points connected by straight lines (polygons), and lines similarly (polylines). The second is the field view. In this conceptualization, geographic variation is characterized by the continuous variation of a