Concepts and Theories of GIS in Business
Peter Keenan, University College Dublin, Ireland
Abstract
This chapter looks at the concepts and theories underlying the application of GIS in business. It discusses the role of information technology in business generally and how GIS is related to other business systems. Different views of GIS use are introduced and the chapter suggests that decision support applications of GIS are more relevant to most businesses than purely operational applications. Porter’s value chain approach is used to assess the potential of GIS to contribute to management. GIS is seen as an emerging technology that will increase importance in business in the future.
Introduction
Information technology (IT) has had a powerful impact on the business world in the last 50 years. IT has facilitated the transformation of business and has allowed new business forms to come into existence. This transformation has reflected the potential of IT both as a cost saving mechanism and as a tool for supporting business decision-making. New developments such as the Internet and mobile applications have an important ongoing impact on business, continuing the process of transformation started by the punched card 50 years earlier. Geographic information systems (GISs) are an area of IT application with a significantly different history from other types of information system. GIS-based applications are now becoming widespread in business, playing a role that reflects both the similarity of GIS to other forms of IT and the distinct characteristics of spatial applications.
2 Keenan
Origins of Geographic and Business Use of IT
Business use of information technology started in the 1950s in payroll, billing and invoice processing applications. These applications exploited data processing techniques that had been previously used by government agencies such as the U.S. Census Bureau. GIS has its origins in the use of IT for geographic related activities in North America in the same period. These early applications were typically government orientated, such as transport planning in Detroit and Chicago and the Canada Geographic Information System (CGIS) (Coppock & Rhind, 1991).
Early business applications of IT employed relatively simple processing that could be automated using the comparatively crude computer technology of the period. One example was payroll processing, where only four or five simple calculations were required for each individual. This computerization of simple numeric processing was an automa- tion of clerical work, analogous to the automation of manufacturing in the earlier part of the 20th century. The high cost of computing in this period meant that this type of application was mainly confined to large organizations with a high volume of transac- tions. While these early data processing applications were relatively unsophisticated, they had a significant impact as they concerned activities critical to business. Data processing techniques allowed these critical operations to be performed faster, more accurately and, above all, more cheaply than manual methods. Despite the relatively high cost of computing at this time, significant cost reductions could be achieved by this automation of the clerical processes required for the day-to-day operation of all businesses. Consequently, early business applications of IT had a widespread impact on routine accounting operations, but were initially much less important in other departments of the organization. In a similar way, the early applications of geographic computer processing were only of interest to the small number of companies involved in map-making, surveying or similar geography-based activities. For example, in the oil industry GIS had a role in exploration at an early stage, but would not have been used in marketing in this sector until much more recently. Many early private sector organi- zations provided consultancy services or GIS software to the public sector. One example would be Tomlinson Associates, set up in 1977 in Ottawa, Canada by Roger Tomlinson, one of the pioneers of GIS. Another example of an early GIS commercial organization would be the Environmental Systems Research Institute established in 1969. This later became ESRI, which is now the main player in the GIS software market.
Development of IT Towards Decision Applications
As IT became more capable and less expensive, business use of computing moved from the automation of clerical processes to decision support applications. This change exploited the superior interaction made possible by time-sharing computers, and the developments in data organization made possible by developments in database manage- ment software. The data available in organizations was initially used to produce regular reports in the form of a Management Information System (MIS). The introduction of improved user interfaces in the 1970s facilitated the introduction of Decision Support
Concepts and Theories of GIS in Business 3
Systems (DSSs). These systems constitute a flexible user-friendly interface linked to problem databases and specific models. As the name suggests, DSSs aim to support, rather than replace, the decision-maker (Sprague, 1980). This form of IT became of interest to managers throughout the organization, as these systems could support decision- making in diverse business functions such as marketing or human resource planning. IT use therefore began to spread throughout all of the business functions, a trend facilitated by the introduction of user-friendly personal computers in the 1980s. Improved network- ing allowed these machines to be connected together, and this has allowed access from a variety of applications to centralized resources such as databases. Modern business applications continue to exploit the rapidly increasing computational power of the computer; but also derive increasing benefits from the ability of IT to store and organize data (databases), distribute the information derived (networking), and present that information in an interactive format (interfaces). This trend also found expression in the development of systems such as Executive Information Systems (EIS) that provide executive management with an overview of business activity within the organization and of competitive forces on the outside.
A similar sequence of developments occurred within GIS, although largely indepen- dently from other forms of IT. The distinct development of GIS was partially a conse- quence of the much larger amounts of data required for spatial applications when compared to business data processing. This meant that the evolution from automation applications to decision support applications was delayed by 10 to 15 years for GIS when compared to traditional business systems (Densham, 1991). Nevertheless, as computer technology became more powerful, the functionality of GIS software greatly increased.
This trend, combined with the lower cost of GIS hardware, has facilitated more ambitious spatial applications. Modern GIS provides distinctive database techniques, specialized data processing and a sophisticated interface for dealing with spatial data.
Consequently, interest in decision support in the GIS field grew in the 1980s when the concept of a Spatial Decision Support System (SDSS) was introduced (Armstrong, Densham, & Rushton, 1986). SDSS was built around GIS with the inclusion of appropriate decision models. By the end of the 1980s, SDSS was a recognized area within the GIS community (Densham, 1991). Over time, decision support applications have found increasing acceptance as an application of GIS and spatial applications have come to constitute an increasing proportion of DSS applications (Keenan, 2003). These applica- tions typically require the synthesis of spatial techniques with other business orientated decision-making approaches based on accounting, financial or operations research techniques.
Initially GIS software was run on mainframe computers, then on relatively expensive graphics workstations. However, as computer performance improved in the 1990s, it has become possible to run GIS software on standard personal computers. This meant that the machines commonly used in businesses were sufficiently powerful to do some useful work with spatial data. Powerful GIS software is now readily available on the Microsoft Windows platform, which is widely used in business and is familiar to business users.
GIS vendors have also recognized the market potential of business applications and GIS software has evolved to meet the needs of this broader set of users, facilitating the design of business applications.
4 Keenan
These developments show a clear trend. Early applications of IT had a cost reduction role, similar to other forms of mass production. However, it was quickly realized that computer technology has a dual nature: it can be used to automate, but as a by-product of this automation it can also produce large amounts of information about the process being automated. In a widely cited book, Zuboff (1988) coins the term informate to describe the ability of technology to provide information about processes as well as automating them.
GIS has also been seen as an informating technology (Madon & Sahay, 1997; Snellen, 2001), as it moves from data processing applications to decision oriented applications.
The informating role of GIS is particularly evident in a business context, where decision- makers value the problem visualization provided by a map, rather than the map itself.
Within the GIS research community, there has been ongoing debate whether GIS is just another information system or whether it has unique characteristics that separate it from other systems. Maguire (1991) conducts a review of the definitions of GIS and suggests that GIS can be seen as a form of IS, with a distinctive orientation towards spatial data and processing. Maguire identifies three views of GIS, with each view focusing on one functional aspect of GIS technology. The map view sees GIS as a map processing or display system. The database view is concerned with simple analysis, such as overlay- ing, buffering. The spatial analysis view focuses on more complex analytical functions such as modeling and decision-making. While these views have something in common with the use of IS for data processing, database management and more elaborate DSS applications, there are also some differences. The map view of GIS includes techniques not widely used in business applications, such as map production using raster opera- tions. The distinction between a map view and database view of a GIS is less clear in Table 1. Computerized Support for Decision Making (Adapted from Turban and Aronson, 2001, pg. 22)
Phase Description Traditional Tools Spatial Tools Early Compute, “crunch numbers,”
summarize, organize
Early computer programs, management science models 1950s - 1960s
Computerized cartography 1960s - 1970s
Intermediate Find, organize and display decision relevant information
Database management systems, MIS 1970s
Workstation GIS 1980s
Current Perform decision relevant computations on decision relevant information; organize and display the results. Query based and user-friendly approach. “What if” analysis
Financial models, spreadsheets, trend exploration, operations research models, decision support systems.
1980s - 1990s
Spatial decision support systems
1990s
Just beginning
Complex and fuzzy decision situations, expanding to collaborative decision making and machine learning
Group support systems, neural computing, knowledge management, fuzzy logic, intelligent agents
Group SDSS, Intelligent spatial interfaces, evolutionary techniques for spatial problems, Geolibraries
Concepts and Theories of GIS in Business 5
business mapping, as these applications generally involve at least a simple database structure to allow the storage of attribute data in addition to geographic data. The spatial analysis view of a GIS implies that the GIS provides models providing analysis of interest to a decision maker. In the business context, appropriate analysis usually requires the addition of specific business models. In this case the GIS is a platform which can be developed into an analysis system with the addition of appropriate models (Hess, Rubin,
& West, 2004; Keenan, 1996). Nevertheless, the development of GIS can be seen as approximating to the phases of development of other forms of IS (Table 2). Presentation mapping, although much more sophisticated, can be related to the fixed format reporting of MIS. The database view of GIS, which allows onscreen query, can be compared to modern EIS systems.
Spatial Visualization
The vast majority of modern GIS applications are characterized by sophisticated graph- ics, and this capacity for visualization allows GIS to provide effective support for problem representation in spatial problems. Long before computer technology was introduced, users gained an improved understanding of spatial problems by the use of maps. While maps were usually initially prepared by governments for political or military reasons, these could also be used for business applications. An important early map, the 1815 geological map of England by Smith (Winchester, 2002), also facilitated business projects such as coal mining and canal construction. In the same period, British Admiralty charts were also seen as an important advantage for British merchant ships trading in distant parts of the world. Early government maps could also be used to assess business potential; one example of this was the 19th century “Atlas to accompany the second report of the Irish Railway Commissioners,” which showed population, traffic flow, geology, and topography all displayed on the same map (Gardner, Griffith, Harness, & Larcom, 1838). This allowed easy understanding of the feasibility of proposed railway routes planned by the private railway companies of that period.
Table 2. Views of GIS
GIS View (Maguire, 1991)
GIS operations Comparable IS System
Map creation Data processing
Map
Map presentation MIS
Business Graphics
Database Simple analysis
Visualization
EIS
Spatial analysis Specialized analysis DSS
6 Keenan
The growth of IS has seen the introduction of new information representation paradigms.
As technology has advanced, users’ ability to work with information has been enhanced by innovations such as graphical user interfaces. Even relatively simple concepts, such as the representation of multiple spreadsheet tables as tabbed worksheets, or the use of hypertext have greatly enhanced the usability of computer systems. The rapid pace of change in technology has provided scope for the use of new problem representations.
However, it takes some time for interface design to take advantage of these develop- ments, as suitable references must be found to assist in the design of new information representation paradigms.
One of the most important strategies in interface design is the use of a visual problem representation to improve user interaction. The area of visual modeling (Bell, 1994) is a recognized part of management support systems. Visual modeling is based on the concept that it is easier to interact with a visual representation of a model than its mathematical equivalent. Geographical techniques have been identified as being rel- evant to the general field of computer graphics, which has had an important influence on business use of IT for decision-making by facilitating visualization applications. Re- searchers from the IS tradition have noted that computer technology is especially appropriate for the display of mapping data. Ives (1982) suggested that maps were too difficult to produce manually for most business applications, and that computerized techniques would make this form of representation much more widely available. Cartog- raphy has been seen as being an important source of principles for the design of business graphics (DeSanctis, 1984); this reflects the fact that many decision makers are accus- tomed to using maps, although this may not be true in all cultures (Sahay & Walsham, 1996; Walsham & Sahay, 1999). Speier (2003) noted that information visualization techniques have been widely applied in science and geography, but have only been recently integrated into business applications. Tegarden (1999) uses the example of the 1854 map of the incidence of Cholera by John Snow to illustrate the power of visualization.
This map is frequently cited as the ancestor of computerized GIS.
As decision-makers in many business sectors are used to the concept of a map, the display of onscreen maps has long been incorporated in computer-based DSS and EIS systems. Many areas of DSS application are concerned with geographic data, an influential early example being the Geodata Analysis and Display system (GADS) (Grace, 1977). GADS was used to build a DSS for the planning of patrol areas for the police department in San Jose, California. This system allowed a police officer to display a map outline and to call up data by geographical zone, showing police calls for service, activity levels, service time, etc. The increasingly widespread use in business of GIS-based systems for map creation and display since GADS reflects the importance of visualization in human information processing.
In the business context, visualization in GIS poses a challenge to interface designers to provide facilities that meet the problem representation needs of users, while also providing convenient ways of interacting with that representation. Computer interface design generally has yet to take full advantage of the increased power of computing and the richer set of possibilities that this offers for user interaction. The complex nature of spatial data requires GIS to use sophisticated visualization techniques to represent information. It is therefore quite challenging for GIS to also to provide an interactive
Concepts and Theories of GIS in Business 7
interface on the same screen. Consequently, GIS applications can especially benefit from better designed human-computer interfaces which meet their specific needs (Hearnshaw
& Medyckyj-Scott, 1993). Visualization has been recognized in the GIS community as an important aspect of GIS (Buttenfield & Mackaness, 1991). This may reflect support for the map view of GIS. One limitation of GIS interface designs is that they are seen to provide a means for visualizing results only, rather than providing a comprehensive problem representation for all stages of the problem (Blaser, Sester, & Egenhofer, 2000). A more comprehensive system would allow problem specification using interactive techniques.
One example is the Tolomeo system (Angehrn & Lüthi, 1990; Angehrn, 1991). In this case, the user can sketch their problem in a geographical context and the Tolomeo system will try to infer the appropriate management science model to use to solve the problem outlined by the user. Another example of sketching might be a real estate agent who could use a GIS interface capable of interpreting a sketch of a customer’s preferences for location (Blaser et al., 2000). In this case the system might interpret the districts where the customer wanted to live and whether they wanted to be close to the sea or other features.
Views of GIS Use
Spatial Data
The spread of GIS technology has been accompanied by simultaneous growth in the amount of digital data available. Extensive collections of spatial data now exist for most developed countries. The same geographical data sets may be used by many different organizations, as many businesses will operate in the same geographic region. Most of the data used in traditional IT applications is sourced within an organization and concerns customers, suppliers, employees, etc. Data of interest in a GIS may include information on existing customers, but will also include data on shared transportation networks and demographic data on people who are not yet customers. Consequently, GIS is somewhat unusual when compared to other business IT applications, in that many users typically outsource both their software and a large part of their data. As the business use of IT moves from internal data processing applications to EIS applications, external data is of increasing importance, and this needs to be effectively linked to external GIS data. The availability and pricing of spatial data is an important factor in the widespread use of GIS, as a significant amount of geographic data is sourced outside the organizations using it.
Geographic data may be collected by the government and made available at little or no cost to organizations that want to use it; this is the case in the U.S. On the other hand, European governments generally seek to recover the cost of spatial data collection from users. Any assessment of the potential of the GIS field to business must take account of the cost and availability of the common data, as well as software and hardware.