Internet GIS for Malaysian Population Analysis

ABSTRACT

    To understand population characteristics in Malaysia, an Internet GIS(Web GIS or Web-based GIS) was created that can analyze demographic statistics, and its applicability was investigated in this paper. This system would be available for foreign researchers, administrators and business people who want geographical information on Malaysia. Users can access a wide variety of attribute data and political map data through the Internet. These are linked on a network, allowing all types of thematic maps to be compiled and statistical processing to be conducted. This system creates choropleth maps that have the functions of "Zoom In", "Zoom Out," graph displays, attribute query, and basic spatial analysis, among other features.

INTRODUCTION

    The purpose of this study is to construct a Geographic Information System on the Internet that has the capability to spatially analyze population characteristics in developing countries (see Fig. 1).

Figure 1.  Schematic depiction of Internet GIS.

    Digitization of statistics, maps, etc., in developing countries is lagging that in developed countries, and there is little opportunity to access digital regional statistics and thematic maps from developing countries. It is nearly impossible to get spatial data through the Internet, and most geographic information is still obtainable only through printed media. Therefore, Internet GIS would be an invaluable resource for geographical research of developing countries as it would allow anyone in the world to use a Web browser to freely create statistical maps and conduct spatial analysis.

    Given this background, this paper will examine the case of Malaysia. To understand the demographics of the country, an Internet GIS was created that can analyze demographic statistics, and its applicability was investigated. This new program should be useful for persons who are interested in the characteristics of Malaysian regions and/or want geographical information on Malaysia, especially foreign researchers, administrators and business people.

MAKE-UP OF THE INTERNET

    Three methods were considered to bring Internet GIS to fruition: 1) processing done mostly by the server, 2) processing done mostly by the user, and 3) processing split between server and user, according to the capabilities of each. To create a highly utile system in which the user can give frequent processing commands in the present Internet GIS, method 2 was chosen since it allows relatively easy interactive spatial analysis.

    The author has opened an experimental web page from his laboratory entitled "Internet GIS for Historical Statistics" (http://giswin.geo.tsukuba.ac.jp/teacher/murayama/index-e.htm) which provides the user with an understanding of the socioeconomic structure of modern Japan, from the early Meiji Era (ca. 1878) to the early Showa Era (ca. 1930). The present study has improved on this experimental system by increasing processing speed and giving it more functionality for spatial analysis in order to create an Internet GIS for demographic statistics of Malaysia.

    It should be noted that the map files are read by a modified version of Java Shapefile Reader Ver. 1.0b, a freeware developed by the Geography Department of Leeds University in the United Kingdom.

    Various services using Internet GIS have been provided to the world since the latter half of the 1990s (Plewe, 1997). Recently it has become much easier to use Java language, which works with any computer operating system, spurring the development of Internet GIS.

    The Geography Department of the University of London is one of the leading institutions involved with Internet GIS research. The Department has organized the Great Britain Historical GIS Program to utilize Internet GIS to analyze and process historical statistics of the UK. Their Web page can be accessed at http://www.geog.qmw.ac.uk/hgis/

    This British project involves data from the late 1830s, when modern statistical data were first collected, to the 1970s. Political (election) and socioeconomic data were linked via computer to allow users to easily compile statistical maps, then the system was made accessible to Internet users worldwide.

    In Japan as well, great strides are being made in creating a practical Internet GIS. For example, there is the "Map on Demand" developed jointly by the DNP and Pasco corporations (http://www.map-ondemand.ne.jp/visitor/index.htm). This system was designed for area marketing and is provided through Java applets. If the user has a Web browser, he or she does not need any special hardware or software. Using this system, it is possible to create map displays and conduct regional analyses on 1,447 statistical categories in the Japanese census. The five types of areal units include two types of local administrative districts (oaza, cho-cho), statistics for chikuchoson divisions, and regional tertiary and quaternary grid statistics. Users can freely choose the scale of map displays.

USER REQUIREMENTS

Hardware

IBM-compatible PC (recommended)

Memory: At least 64 megabytes

Hard disk memory: At least 2 gigabytes

Software

Windows 95/98/NT4.0

JDK 1.1.X or above

Netscape 4.0.6 or above, or

Microsoft Internet Explorer 4.0.1 or above

SYSTEM STRUCTURE

Server

Computer: Sun Workstation

OS: UNIX

HTTP: NCSA

File structure

Maps: Map file and associated data

Metadata: File locational information, names of attribute data categories

DATA

    The object of the present research is Peninsular Malaysia. The data used here were taken from the "Population and Housing Census of Malaysia"¹⁾ compiled by the Department of Statistics Malaysia. Population statistics from 1911 on were available for Peninsular Malaysia.²⁾

    There was a total of 188 variables analyzed in the present study, representing the statistical categories of sex, religion, school attendance, citizenship, status in labour force, employment, occupation, possession of household appliances, housing units, material of outer walls, age of living quarters, occupancy, type of lighting, type of toilet facility, total population, highest education certificate, and hours worked (see Table 1). Regarding the spatial scale, three different administrative units have been used here: State, Administrative District, and Mukim (the smallest unit). In Peninsular Malaysia there are 12 states (including the federal territory of Kuala Lumpur), 81 administrative districts, and 844 mukims.

    In addition to Peninsular Malaysia, the present study also used data from Kuala Lumpur, especially the very minute statistical districts to construct the Internet GIS. The reason for this inclusion is that since Kuala Lumpur is a major city of 1.23 million inhabitants (1991 data), it was determined that users would have a very high need for choropleth maps and spatial analysis of this area.³⁾

    The statistical categories for 1980 are shown in Table 2; for 1991, in Table 3.

    The above regional statistics were compiled into a database in the form of a geographical matrix and are stored on the server at the University of Tsukuba. It should be noted that all values are expressed as percentages of total values. For example, the value for Male (under "Sex") in Table 1 is the proportion of males to the entire population.

DESCRIPTION OF THE SYSTEM

    This system creates choropleth maps that have the functions of "Zoom In", "Zoom Out," graph displays, attribute query, and basic spatial analysis, among other features. Users can access a wide variety of attribute data and political map data through the Internet. These are linked on a network, allowing all types of choropleth maps to be compiled and statistical processing to be conducted (see Fig. 2).

 Fig. 2.  Home Page.

    First, the user selects the spatial unit to be analyzed, i.e., state, administrative district, or mukim. Figures 3 and 4 show sample screens for state and mukim, respectively.

Fig. 3.  Selection of State unit.

Fig. 4.  Selection of Mukim unit.

Creation of choropleth maps

    The user then selects a blank map from the pull down menu to depict the attributes and their distribution on the choropleth map. Depending on the needs of the user, the legend can be divided into 2 to 7 hierarchical levels, and changed at will.

    These hierarchical levels can be selected using three different modes: 1) "equal size intervals", which make all hierarchical intervals equal in size; 2) "equal number of features", which makes the number of features the same in all levels; and 3) "standard deviation", which shows mean attribute values as 0 and variance as 1.

    Figure 5 shows a sample choropleth map in which the distribution of the male ratio in Peninsular Malaysia in 1991 is shown with 7 hierarchical levels in "equal size interval" mode. Figure 6 uses the same levels and mode to depict "Ethnic group: Chinese ratio" for Kuala Lumpur for 1991.

Fig. 5.  Map display of 1991 Male ratio in Peninsular Malaysia based on equal size intervals.

Fig. 6.  Map display of 1991 Chinese ratio in Kuala Lumpur based on equal size intervals.

Information (Info) display

    On the screen displaying the map, the user, simply by clicking the mouse on a region, can obtain information on it such as code number, attribute value, and place names.

Labels

    Clicking the "Label Display" button on the tool bar will automatically display place names.

Zoom in, Zoom Out, and Pan

    The enlarging and reduction functions ("Zoom In" and "Zoom Out") make it possible to change the scale of the map. To enlarge an area, the user clicks the "Zoom In" button on the tool bar, then designates the area to be enlarged with the mouse. Thereupon, the mouse cursor will turn into a "+" sign, and the designated area of the map will be displayed on a larger scale. Figure 7 shows an example of an enlarged section of a choropleth map for the Islam ratio of the Selangor region.

Fig. 7.  Zoom In function (Islam ratio in the Selangor region, 1991).

    To reduce the size of the displayed area, the user clicks the "Zoom Out" button on the tool bar. The mouse cursor will then turn into a "+" sign, and the designated area of the map will be displayed on a smaller scale.

    By clicking the "Pan" button, the user can freely move the position of the map on the screen by holding down the mouse button, dragging the map to the desired location, then releasing the mouse button.

    Clicking the "Initial Size" button at the top of the screen will return the map to its original scale and position.

Attribute query

    This is a function that searches attribute values based on numerical parameters and displays areas having those values on the map. For example, by setting the population parameter as "10,000 or over", regions of less than 10,000 population will not be displayed. Figure 8 shows an example of regions at the mukim level for Peninsular Malaysia in 1991 in which the Islamic ratio (proportion of Moslems to entire population) is 0.6 (60%) or greater.

Fig. 8.  Attribute query function.

Graph display

    By using this function, the relation between two variables can be displayed as a scattergram. Figure 9 shows Islam ratio on the x-axis, and agricultural worker ratio on the y-axis to show their relation in graphic form.

Fig. 9.  Sample scattergram.

Analysis

    This function allows the user to calculate statistics (mean, standard deviation, variance) for selected variables. Figure 10 shows an example of Islam ratio for all 844 mukims in Peninsular Malaysia in which the mean is 0.738, the standard deviation is 0.258 and the variance is 0.066.

Fig. 10.  Analysis function.

    This function allows users to combine a map with a graph to find patterns, relationships, etc.  For example, clicking a point (region) on a distribution map will make that region appear on a regular map, while clicking a region on the regular map will make it appear as a mark on the distribution map.

CONCLUDING REMARKS

    The system in the present study has been constructed with both English and Japanese versions. This allows it to be used in Japanese classroom education, such as in high school geography classes.

    Hereafter, it will be very important to gradually expand the database by digitizing data from other years of the Population and Housing Census of Malaysia to provide a better temporal understanding of these data. In addition, it will be necessary to add spatial analytical functions to meet the high demands on Internet GIS.

    This system, presently in the experimental stage, is accessible over the Internet at ../../index-e.html

Acknowledgment

    I would like to express my appreciation to Mr. Hisaji Ono of the Pasco Corporation for creating this Internet GIS program.

Notes

1) The Census of 1991 covered all persons including foreigners living or who were in Malaysia on Census Day, i.e., 14 August 1991. However, Malaysian citizens and permanent residents who were temporarily away from the country on Census Day because of duties, touring or studying were excluded. The 1991 Census also covered wayfarers. The category wayfarers refer to persons who, on Census Day (between 12 midnight, 13 August 1991 and 7 a.m., 14 August 1991), were travelling along highways or were on trains or boats (ships). Wayfarers were included in the place in which they were found at that particular time（DSM, 1992.

2) Data for some regions in Peninsular Malaysia are available from 1871, but they do not cover the entire region. Since 1911, population censuses have been taken in 1921, 1931, 1947, 1957, 1970, 1980 and 1991. This paper looks at only the data from 1980 and 1991, but in the future an effort will be made to eventually create a chronological database.

3) The number of basic units used in the present study was 642 for 1980, and 879 for 1991.

References

DSM (1992): Preliminary count report for urban and rural areas. Department of Statistics Malaysia,, Kuala Lumpur, p.10.

Murayama, Y. and H. Ono (1998): Development of Internet GIS: Map information system of regional statistics in the Meiji period. Tsukuba Studies in Human Geography, 23, 99-128.

Murayama, Y. (1999): Internet GIS: Map information systems of population census in 1920 and 1930. Tsukuba Studies in Human Geography, 24, 59-79.

Plewe, B. (1997):GIS online: information retrieval, mapping, and the internet. OnWord Press, Santa Fe, 311p.

Home Page