GIS Prototyping: Evaluating Representation Models for Spatial Data

D. Pfoser
Master Thesis
MT9502 (September, 1995)
Supervised by
o. Univ.-Prof. Dr. Michael Schrefl
Instructed by
Prof. Dr. Jan Paredaens
Accomplished at
University Linz, Institute of Business Informatics - Data & Knowledge Engineering

Abstract (English)

Geographical Information Systems, GIS, are a recent development in computer science. The goal of GIS, we might also call them spatial databases, is to manipulate spatial information and make useful information from the geographic data stored in the database, readily available.

Solving the twofold problem of representing and manipulating spatial information is the aim of this work.

The first part of this work studies existing theoretical models. previous works describe a way to model spatial information by means of elemental spatial entities in the Euclidean Space. Further discussion is about topological relationships which are a subset of a given set of operators defined on elemental spatial entities.

In later chapters the focus is on representing spatial data in a given database system. The models described are the result of experimental prototyping. For this process I used the DBMS Postgres and a graphical frontend application, CEO++, to display the results of query processing.

Spatial data used in this process were obtained from the Tiger/Line Census dataset.

The two described models represent two extremes. The database containing spatial information modeled by means of a graph is a traditional system whereas spatial data types are a semantical extension of the database. The purpose of this work is to bridge the gap between existing theoretical concepts for dealing with spatial information and a practical approach.

Why make such an effort to display some cheesy lines? Why use a system software like Postgres which is on the one hand full of bugs, and on the other hand tricky to install and run?

To obtain everything from nothing, to group fancy numbers to valuable geographic information?

Good question.

This whole enterprise started as a part of a bigger project, MAGNUM. Though it was not planned in the beginning, my little part in the bigger game became more and more independent. Of course there is still relevance to the original goal of designing a prototype for a geographic information system, GIS, but the longer the work lasted, the larger became my personal need to know more about GIS in general. My interest in the underlying mathematical theory as well as the current state of the art of GIS research grew.

This demand also is not temporary. In the long term, this topic also might serve as a basis for my Ph.D. thesis.

Things, and therefore interests change. This could be a summarizing statement for research in general. In particular, it was an experience I encountered during the work on this thesis. Never be afraid of confessing a change in your attitudes, whether they are in work or in life. Let them influence your personal evolution and never be afraid of them.