Academic journal article Geographical Analysis

SANET: A Toolbox for Spatial Analysis on a Network

Academic journal article Geographical Analysis

SANET: A Toolbox for Spatial Analysis on a Network

Article excerpt

This article shows a geographical information systems (GIS)-based toolbox for analyzing spatial phenomena that occur on a network (e.g., traffic accidents) or almost along a network (e.g., fast-food stores in a downtown). The toolbox contains 13 tools: random point generation on a network, the Voronoi diagram, the K-function and cross K-function methods, the unconditional and conditional nearest-neighbor distance methods, the Hull model, and preprocessing tools. The article also shows a few actual analyses carried out with these tools.

Introduction

This article outlines the tools of a geographical information systems (GIS)-based toolbox, called SANET (Spatial Analysis on a NETwork), which is designed to analyze spatial phenomena that occur on networks (referred to as network spatial phenomena). SANET can deal with two types of network spatial phenomena. The first type is a class of phenomena that occurs on a network (i.e., physically located on a network). A typical example is traffic accidents. The second type is a class of phenomena that occurs "almost" along a network (i.e., not physically located on a network). A typical example is the distribution of retail stores in an urbanized area. An actual distribution is shown in Fig. 1(a) where the points represent the centroids of noodle restaurants in Shibuya, Tokyo. In fact, the entrances of these stores face streets (Fig. 1(b)).

In the literature, network spatial phenomena are usually analyzed by spatial methods assuming that: the real world is represented by a plane, events or facilities are represented by points on the plane, and the distance between two points is measured by the Euclidean distance. We call these methods planar spatial methods and spatial analysis with such methods planar spatial analysis. If we apply planar spatial methods to network spatial phenomena like those shown in Fig. 1, we are likely to reach false conclusions. Actually, Yamada and Thill (2004) illustrated this pitfall with traffic accident data in Buffalo (also see Okabe et al. 2006).

[FIGURE 1 OMITTED]

To avoid such false conclusions and to analyze network spatial phenomena properly, we should use spatial methods that assume that: the real world is represented by a network embedded on a plane, events or facilities (their entrances) are represented by points on the network, and the distance between two points is measured by the shortest-path distance. We call such methods network spatial methods and spatial analysis with these methods network spatial analysis.

Although network spatial methods are more appropriate than planar spatial methods to analyze network spatial phenomena, the former methods give tougher computation problems. Because of this difficulty, researchers tend to use planar spatial methods, presuming that the results obtained from planar spatial methods are approximately the same as those obtained from network spatial methods. As mentioned above, however, this presumption is likely to be false. To analyze network spatial phenomena properly, we must have easy tools. SANET meets this need by providing user-friendly tools for network spatial analysis.

Components of SANET

SANET consists of two components: the main SANET program and the SANET-ArcView interface. The main SANET program is written in C++, and computes the equations and functions used in network spatial methods in SANET. The SANET-ArcView interface imports data from ArcView to the main SANET program, and exports results from the main SANET program to ArcView.

We make two remarks here. First, the main SANET program is almost independent of the GIS viewer used, and it can communicate with any GIS viewer through standard import and export files. If a user is familiar with programming, development of an interface between the main SANET program and any preferred GIS viewer is not difficult. Second, SANET is free for nonprofit use. The first version was released in 2002, and at the time of writing (March 2004), the second version (Okabe, Okunuki, and Shiode 2004) has become available and can be downloaded from http://okabe. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.