A Computer System for Mapping And Analyzing Transportation Networks

A COMPUTER SYSTEM FOR MAPPING AND ANALYZING TRANSPORTATION NETWORKS Robert I. Wittick* In recent years transportation geography has become heavily dependent on analytical methodology. This dependence has brought with it a need for computational aids in the use of these methods. For example, while a set of graph theoretic measures for a network can be derived manually, the efficient solution of a large linear programming transportation problem requires the use of a computer. It is likewise necessary to solve most allocation models and network assignment models with the help of a computer. Although computer programs now exist for most of this analytical methodology, they are often difficult to acquire and occasionally unreliable . While it would be feasible to upgrade these existing programs , it is more logical to integrate the algorithms via a systems approach into a cohesive package of user oriented, computer efficient, well documented, and easily exportable software. In other words, transportation geography needs a computer system that can both aid the researcher in his data analysis and give the student of transportation geography practical experience in the use of those analytical methods discussed in class. This paper discusses the development and design of a computer system called FLOW, an integrated software package for mapping and analyzing transportation networks. The paper first describes the rather unique mapping capabilities of the system, followed by a discussion of the integration of commonly used transportation geography methods into the system structure. A perusal of the current textbooks in transportation geography supports the statement that modern transportation geography is heavily analytical. Unfortunately, students frequently have difficulty understanding the application of the methods when they are not given * Dr. Wittick is assistant professor of geography and a member of the Computer Institute for Social Science Research at Michigan State University. This paper was accepted for publication in February, 1976. Vol. XVI, No. 1 75 the opportunity to use them. Because most analyses used in transportation are so complex that use of computers is required, it is necessary that students in the field be given practical experience in computer methods. This should be done in a way which minimizes time spent teaching the mechanics of computer usage; otherwise, applications may be obscured by emphasis upon techniques. In designing a system such as FLOW, it was necessary to keep several goals in mind. 1) The system had to be easy to use. This is particularly important when the user is a typical student in transportation geography who has had no previous computer exposure. 2) The system had to be flexible in operation, providing the researcher with numerous analytic and graphic techniques applicable to spatial flows and networks. Although these first two goals are somewhat contradictory , i.e., designing one system for both the sophisticated researcher and novice student, both goals were considered important and should be achieved if possible. 3) The computer system had to be efficient in the utilization of computer resources. This is particularly important when the system is to be used by a large number of students. 4) The system had to be general in structure and easily exportable to other installations. In essence, the overriding goal of the project was to design a computer system which could easily be used by students and by experienced researchers in transportation geography for analyzing and mapping transportation networks and flows. DEVELOPMENT OF MAPPING SOFTWARE. In 1973 a project was begun within the Computer Institute for Social Science Research to develop a computer program which would graphically portray quantitative flow data using a digital plotter. Up to this time there were numerous computer programs in geography for mapping discrete and continuous point data and area data; however, there was little in the way of automated mapping software for quantitative flow data. After numerous experiments with various graphic techniques, it was decided that two methods were best suited for automatically portraying quantitative information along flows or links. The first consisted of the construction of constant width corridors connecting the nodes along the link with shading gradients drawn inside the corridors (Figure 1). The shading gradients varied by class intervals with 76 Southeastern Geographer FLOH INTERVALS ? ? ? ? 1 .50 _ 2.89 2.90 _ 4.29 4.30...