c
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B6. Istanbul 2004
load time and the increasing requirements to the client
computer. The representation of more than 1.000 municipalities
in Mecklenburg-Western Pomerania could not be achieved
within a reasonable time. These performance issues limited the
representation on the borough level in the administrative
hierarchy in Germany. On the client side a version of the Java
Runtime environment is needed.
A further open environment for the interactive visualization of
spatial data is SVG. SVG (Scalable Vector Graphics) is a
language to describe two-dimensional graphics by XML.
Already different standardisations are given by the W3C and it
can be assumed that SVG intersperses as a standard in all future
WWW-viewers. Highly developed interactive SVG applications
are converted with the help of a supplementing script language.
The full access to the SVG Document Object Model (DOM) is
possible e.g. with JavaScript. Thus at the same time access to
all further XHTML and SVG elements in the same web page is
given. SVG applications are quality-free scalable. The file size
is small using text compression and the loading time is short.
SVG is a kind of XML and is based on pure ASCII files. These
files can be built dynamically with PHP and be embedded in the
Web application. Hereby the spatial data are stored in the open
standard GML (Geographic Markup Language) in the MySQL
database and visualised depending upon interaction. The
principle of the open SVG map server according to
www.carto.net is used.
For the computation of statistical measures the language R is
used. R is a development of the University of Auckland and a
computer language for statistic analyses. R can be merged into
different application packages. Via the CGI interface of the
Web server the functions can be called dynamically, R can
access the data in ‘the database directly. For our local
applications the combination of MySQL/PHP and SVG was
suitable and the performance was satisfying.
In the future employment the developed software components
are merged with the respective teach and learning platforms of
the university e.g. WebCT, Ilias, Blackboard, or stud.IP. The
learning-platform offers the basic functionalities such as user
administration and communication tools.
2.2 Target, content and data set in the learning module
The target of the learning unit is to learn about spatial
visualisation and spatial analysis with respect to population
statistics. Subject of population statistics is to apply statistical
methods and procedures for the numerical collection,
representation, analysis and interpretation of the development
on population in a special region. Statisticians know that
diagrams of statistical or calculated data are easier to interpret.
With increasing number of data sets graphical visualisations are
also easier to be interpreted than tables. Particularly for the
representation of spatial distributions of the population map
techniques are very helpful, because spatial conditions and
relations can be better recognized.
National and federal offices for statistic data supply official
population data on different aggregation stages. The smallest
administrative unit, on which the information is published, is
the municipality. A lot of characteristic data are raised annually
on the municipality level at statistical offices. In this teaching
material comparatively the population existence Mecklenburg-
Western Pomerania are graphically represented for the time
stamp 31.12.1990 and 31.12.2000. These data were published
by the statistic national office in the year 2001. In
Mecklenburg-Western Pomerania there were altogether 1203
administrative districts (municipalities, districts, urban areas
and country) in the year 2000. Thus the table generated from
165
that data set with the appropriate population information has
over 1200 lines and according to the acquired population
characteristics many columns. It is very hardly readable. The
following table shows a cut-out of these data from
Mecklenburg-Western Pomerania in the reference to the
population numbers. Each municipality represents a regional
date with its attributes. Each municipality has its specific size,
to which the values refer.
Gemeinde- Bevstd. Bevsid: Bevstd.
2096 60
Table 1: Table excerpt - official statistic data of Mecklenburg-
Western Pomerania
2.3 Asession with the learning module
At the beginning of the learning unit the student tries to conquer
the data set with the acquired statistic knowledge from
preceding chapters. For instance questions such as "find the
largest and smallest municipality, municipality with the highest,
middle and lowest total population" can easily be answered by
using statistical measures such as span width, average value etc.
The student can compute also new data e.g. the population
density as a quotient from population conditions to the surface
size of the administrative unit. Thereby the scholar learns to
differentiate between absolute (e.g. population conditions) and
relative values (e.g. population density). He may compare the
results of different municipalities. As long as these are
identified with their names, the student has no problems,
particularly if he has a certain local knowledge in that region.
But the scholar will recognize that it is not easy to find patterns
on spatial distributions in these data.
For each municipality and/or each district in the table the
geometric borders are stored. This is now linked to the attribute
data of the municipality and/or the district making use of the
primary key, the name of the municipality. The polygons are
individual geometries associated to the appropriate data record.
In the following lesson the student sees the spatial allocation of
the municipalities in the country Mecklenburg-Western
Pomerania. For the first time the scholar makes himself familiar
with different types of representation. Absolute values are
applied for entities like the population existence to the surfaces
to a map. Relative values are used, if statistic proportionality
factors have to be shown such as per cent or values relative to
an uniform base factor (e.g. habitant/km?). For absolute value
representations signatures for point or area elements may be
used, which possess qualitative and quantitative attributes. The
signatures are represented at standard positions related to the
unit of the area. In order to represent several values in the
comparison, one uses diagrams. The values are represented in
addition in different scale, so that the values are recognizable
despite different dimensions (e.g. one point per 1000 person).
