Development of Integrated Spatial Analysis System Using Open Sources Hisaji Ono PASCO Corporation 1-1-2, Higashiyama, Meguro-ku, TOKYO, JAPAN; Telephone: +81 (03)3421 5846 FAX: +81 (03)3421 5846 Email: hi_ono2001@ybb.ne.jp Yuji Murayama Division of Geography, Institute of Geoscience, University of Tsukuba Tsukuba, IBARAKI 305-8571, JAPAN; Telephone +81 (029) 853-4211 FAX +81 (029) 853-4211 Email mura@atm.geo.tsukuba.ac.jp Abstract GeoComputation which attempts to integrate the quantitative geography and spatial visualization technology, has been very popular among GIS communities since the beginning of the 1990s. Especially the construction of platform for spatial analysis is one of the important research topics in the advanced countries. Luc Anselin who made SpaceStat, a spatial analysis software, has played a significant role in the development of GIS, and recently the Centre for Computational Geography at the University of Leeds, GeoVISTA Center at the University of Pennsylvania and Center for Spatially Integrated Social Science in the University of California, Santa Barbara have given a great impact on the development of spatial and regional analysis using GIS. Given this background, this study tries to construct the system of integrated spatial analysis which is useful for the GIS education for geography majors in the university. The functions of this system that was built by using only open sources, include the mapping, spatial search, TIN, overlay, point pattern analysis, spatial autocorrelation, multivariate analysis such as regression analysis, factor analysis and cluster analysis, neural network, tessellation, and so on. This system works at Windows98/Me/NT/2000/XP. This software is now available through the following Internet address(http://land.geo.tsukuba.ac.jp/teacher/murayama/geodbminer/index.html(j apanese)). 1. Introduction GeoComputation which attempts to integrate the quantitative geography and spatial visualization technology, has been very popular among GIS communities since the beginning
of the 1990s. Especially the construction of platform for spatial analysis is one of the important research topics in the advanced countries. Luc Anselin(1993) who made SpaceStat, a spatial analysis software, has played a significant role in the development of GIS, and recently the Centre for Computational Geography at the University of Leeds, GeoVISTA Center at the University of Pennsylvania and Center for Spatially Integrated Social Science in the University of California, Santa Barbara have given a great impact on the development of spatial and regional analysis using GIS. Given this background, this study tries to construct the system of integrated spatial analysis which is useful for the GIS education for geography majors in the university. The functions of this system that was built by using only open sources, include the mapping, spatial search, TIN, overlay, point pattern analysis, spatial autocorrelation, multivariate analysis such as regression analysis, factor analysis and cluster analysis, neural network, tessellation, and so on. This system works at Windows98/Me/NT/2000/XP. 2. System Integrated Spatial Analysis System s architecture is showed in figure 1. This system s GUI part was built by Java. As shown in Figure1, this system s GIS engine employs GeoTools1 for Java with some modifications, abilities of spatial query and spatial analysis enabled and empowered by JTS, Java Topology Suite, API and regional analysis owes to R language with its contributed packages. And this system can process only ArcView shape file formats currently.
Figure 1..System Architecture of Integrated Spatial Analysis System. Software required by Integrated Spatial Analysis System are listed in Table 1. Table 1. Software required by Integrated Spatial Analysis System Software Used Version Java2SE 1.4.1 or later R 1.7.0 or later R(D)COM 1.2
R contributed packages used by Integrated Spatial Analysis System are listed in Table 2. Table 2. R contributed packages used by Integrated Spatial Analysis System Package fields maptools splancs spdep tripack Function Kernel Density Estimation Loading ArcView Shape files Nearest Neighbourhood, K function Spatial Autocorrelation TIN, Voronois, convex hull 3. Functions 3.1 Choropleth Mapping This system supports three types choropleth shading, equal length, quantile and Jenk s natural breaks. And also supports circle dot mapping. 3.2 Gridding Creating grid surface. 3.3 Cartogram This system can create two types of area cartograms, continuous area cartogram(figure 2) and non continuous area cartogram(figure 3).
Figure 2. Continuous Area Cartogram 3.5 Surface This system can create grid surfaces.. Figure 3. Non Continuous Area Cartogram 3.6 Query This system can issue spatial and aspatial queries against shape files.
1) spatial query Spatial query uses JTS API to perform following operators(table 3). Table 3. Operators of spatial query Operator contains contained disjoint equals intersects overlaps touches within Function Select parts of target layer contained by subject layer Select parts of subject layer contained by target layer Select parts of target layer not contained by subject layer Select parts of target layer equalled to subject layer Select parts of target layer intersected subject layer Select parts of target layer overlapped by subject layer Select parts of target layer touched by subject layer Select a part of target layer within subject layer An example of spatial query touches is showed in Figure 4. Figure 4. Example of touches
2) aspatial query 3.7 Spatial Analysis This system can perform following spatial analyses using JTS API. And R s contributed packages. 1) polygon overlay Operators of overlay are listed in Table 4 and polygon operators diagram is showed in Figure 5. Table 4. Operators of polygon overlay Operator intersection difference symdifference union Function clipping target layer by subject layer difference between target and subject layer symdifference between target and subject layer merge of target and subject layer
A B A. Ž ƒ Ž(B) A.union(B) A. e Ž ƒ (B) B. e Ž ƒ (A) A.symDifference(B) Figure 5. diagram of polygon overlay(vivid SOLUTIONS(2002)) An example of polygon overlay is showed in Figure.6.
Figure 6.. Intersection s example(clipped town districts by 500m buffering of stations) 2) Buffering 3) Dissolve 4) TIN 5) voronois diagram 4) convex hull 3.8 Regional Analysis This system can deal with following methods. 1) descriptive statistics 2) multivariate analysis 3) ESDA 5) point pattern analysis 6) ANN(Artificial Neural Network) 3.9 Spatial Autocorrelation Using Byband s spdep(2002), this system can cover almost methods of this area. This system can perform following analysis methods. 1) Global Moran s I 2) Global Geary s c 3) Local Moran 4) Local Geary 5) Join count 6) G & G * statistics Figure 7 shows output example of G * Statistics.
Figure 7 G * Statistics 4. Future Plans We have following plans for further developing this system. (1) Adding more analysis techniques. (2) Adding more visualisation functions. (3) Strengthening functions of saving results by analyses. (4) Fulfilment of printing function. (5).Database connection (6) Strengthening Internet connecting function (7) Scripting for automatic analysis (8) Internationalization (9) Web Applicationization. This software is now available through the following Internet address(http://land.geo.tsukuba.ac.jp/teacher/murayama/geodbminer/index.html (currently Japanese, but soon available in English). 5. References ANSELIN, L., 1993, SpaceStat: A Program for the Statistical Analysis of Spatial Data (Santa Barbara: NCGIA). ANSELIN, L., 1995, Local indicators of spatial association LISA, Geographical Analysis, 27, 93-115 ATKINSON, P. and MARTIN, D., 2000, GIS and Geocomputation (London: Taylor & Francis). BIVAND, R. S. and GEBHARDT, A., 2000, Implementing functions for spatial statistical analysis using the R language, Journal of Geographical Systems, 2, 307-317.
BIVAND, R., S., and Neteler, M., 2000, Opensource geocomputation: using the R data analysis language integrated with GRASS GIS and PostgreSQL data base systems,. Proceedings 5th Conference on GeoComputation, University of Greenwich. BIVAND, R., 2002, The spdep package, http://cran.r-project.org/doc/packages/spatstat.pdf. DORLING, D., 1996, Area cartograms: their use and creation. (Concepts and Techniques in Modern Geography CATMOG), 59. GETIS, A., and ORD, J., K., 1992, The analysis of spatial association by use of distance statistics. Geographical Analysis, 24, 189-199 FOTHERINGHAM, A., S., BRUNSDON, C. and CHARLTON, M., 2000, Quantitative Geography Perspectives on Spatial Data Analysis (London: SAGE Publications). LEE, J., and WONG, W., S., 2000, Statistical Analysis with ArcView GIS( New York: John Wiley & Sons). LONGERY, P., BROOKS, S., M., McDONNELL, R., and MACMILLAN, B., eds. 1998 GeoComputation, Primer (Chichester: John Wiley & Sons). OPENSHAW, S., and ABRAHART, R., J., eds.,2000, GeoComputation(London: Taylor & Francis). ORD, J., K., and GETIS, A., 1995, Local spatial autocorrelation statistics: distributional issues and an application,. Geographical Analysis, 27, 286-296 VIVID SOLUTIONS, 2002, JTS technical specifications, http://www.vividsolutions.com/jts/jtshome.htm..