ePortfolio | GIS Project

GIS Project

Course description

The GIS Project was ment to be a "playground" for any kind of geoinformational problem / challenge and we were completely free to choose any topic with relation to geoinformatics. In a group of two we choose to work on a geodatabase to perform an accessibility analysis for the branches of an Austrian food supply company in Salzburg. All calculations have been performed within the database and a GIS is only used to display the data. Our main focus was put on the excessive use of Open Source Software, such as Quantum GIS, PostgreSQL or the pgRouting-plugin for PostGIS.

What we wanted to do

The overall goal of this project was the calculation and visualization of catchment-areas of Spar branches in the city of Salzburg for three different types of movement (by foot, by bicycle and by car) on the basis of a road-graph. Such an accessibility analysis could serve as a basis for the decision about the location of further branches and give a good overview of the market coverage of a company. Other very important goals are the excessive use of open source software, such as PostGIS or Quantum GIS and the work with object-relational databases. During the semester's course we gained experience in working with those spatial databases and their peculiarities.

Used software

PostgreSQL 8.4

A powerful open source object-relational database system. The database is used to store all geo-related content. This comprises mainly the road network, the Spar markets, the administrative boundary area of Salzburg and tables for the analysis, the point clouds and alpha shape polygons for the catchment areas. The database runs as a server-service in the background and can be accessed through clients like a WebGIS or a Desktop GIS like QuantumGIS.

PostGIS 1.5

Spatially enabling the PostgreSQL server and allowing it to be used for geographic information systems (GIS). PostGIS follows the OpenGIS "Simple Features Specification for SQL". This means that all geo processing tasks can be performed with SQL queries and are stored in the PostgreSQL database in form of tables or views. PostGIS enables the PostgreSQL database to additionally store the geometry type of a feature and his spatial reference system identifier (SRID) as columns in the specific table.

pgAdmin III 1.10

A web-based administration-tool for PostgreSQL. This tools allows a convenient manipulation of the database via a graphical user interface (GUI). You still have to know the SQL commands, but you don't have to bother about issues like database connection strings etc. that normally have to be performed over a command line interface (CLI).

pgRouting 1.03

Extends the PostGIS geospatial database with routing functionality to perform our catchment area analysis. Therefore the driving distance function of pgRouting is used.

Quantum GIS 1.8.0

An open source Desktop GIS with PostGIS database and OpenLayer connection availability to visualize and analyze the database content.

Results

1. Study area

In the picture on the left you can see the administrative boundaries of the city of Salzburg and an Open Street Map road graph as background layer. The green dots represent all Spar branches within the city. We decided to perform our calculations for only one branch, although for three different types of movement (by foot, by bicycle and by car) within the same time-span of five minutes. The expected result is a map that reveals how far you can get within this given time-span by a corresponding type of movement.

2. Point cloud

The next step comprises some calculations within the database to create a network topology. This is necessary because the routing function has to 'know' where all nodes are located. Further all types of roades and paths have a certain speed value assigned. As every road-segment has a certain length and a certain speed value, it's easy to calculate how long it takes to pass one segment. The routing algorithm continues to sum up the road segments until the limit of 5 minutes is reached. The example on the left reveals a catchment-area for 5 minutes by car, starting from the Spar branch in the Rudolf-Biebl-Straße. Every green dot on the map represents one road-node.

3. Interpolation

This interpolation has been calculated within Quantum GIS. It can of course only draw a static picture and traffic jams or construction works have not been taken into account. But still, it delivers quite plausible and correct results. As you can see, the Mönchsberg-mountain is completely spared which makes perfect sense as you would not make it on top of it within 5 minutes from Rudolf-Biebl-Straße. You can also see that the airport-undertake has been taken into account which leads to a bottleneck-like effect.

4. Final result

This map shows our final result. It displays the accessibility of the Spar-branch at Rudolf-Biebl-Straße for the three different types of movement. Accessibility by foot is colored brownish-red, accessibility by bicycle is colored green and the area which is accessible by car within 5 minutes is colored blue. Such a calculation could of course be performed for any place anywhere in the city. We decided to do it for this branch, it is situated in the center of the city.

If you have a closer look at the motorway 'A1' in the north-west, you can see that the area is not coherent, but spotted. This effect occured after the interpolation it we were not quite sure why it happened, however we are very pleased with the result.