Paulo Morgado*, Eduardo Gomes, Nuno Costa
Centre of Geographic Studies, Institute of Geography and Spatial Planning of the University of Lisbon, 1600 Lisboa, Portugal
Competing visions? Simulating alternative coastal futures using a GISANN web application
In this paper, we demonstrate the use of scenario building in the context of contested land use visions.
We examine a small coastal community located 20 kms south of Lisbon. In Almada e Trafaria/Costa da
Caparica, competing stakeholders such as central government, local government, environmental NGO's
and private companies each have competing development visions for the area. These include the
development of recreation and leisure facilities, a container terminal and the re-naturalization of unused
land. We illustrate the added value of the GIS-ANN tool in steering negotiations between these different
visions and the potential of a scenario building web application as a tool for problem solving.
The emergence of user-created GIS-based web content in Planning has transformed passive users and
consumers of geospatial information into active contributors to the development of spatial visions of the
future. It allows stakeholders to gauge alternative future land uses thus making planning and decisionmaking
processes potentially more transparent and democratic. In this paper, we detail a new method
that enhances GIS-web-based public participation. We build on a combination of GIS basic capabilities
and the data mining methods of Artificial Neural Networks (ANN), namely Multilayer Perceptron (MLP)
packaged in a friendly (GUI) user interface that runs on the Google Earth platform. Users will be able to
articulate different spatial development scenarios for a specific area, to conduct sensitivity analyses for various competing scenarios and to explore causal connections between them
To implement a GIS-ANN-MLP model that runs through a Web application we opted for a free open source neural network library, named Fast Artificial Neural Network (FANN). Before implementation, FANN was tested through the FANNTool GUI This enables us to prepare the data in FANN library standard, design, train, test and run the ANN. This enables us to prepare the data in FANN library standard, design, train, test and run the ANN.
Centre of Geographic Studies, Institute of Geography and Spatial Planning of the University of Lisbon, 1600 Lisboa, Portugal
Competing visions? Simulating alternative coastal futures using a GISANN web application
In this paper, we demonstrate the use of scenario building in the context of contested land use visions.
We examine a small coastal community located 20 kms south of Lisbon. In Almada e Trafaria/Costa da
Caparica, competing stakeholders such as central government, local government, environmental NGO's
and private companies each have competing development visions for the area. These include the
development of recreation and leisure facilities, a container terminal and the re-naturalization of unused
land. We illustrate the added value of the GIS-ANN tool in steering negotiations between these different
visions and the potential of a scenario building web application as a tool for problem solving.
The emergence of user-created GIS-based web content in Planning has transformed passive users and
consumers of geospatial information into active contributors to the development of spatial visions of the
future. It allows stakeholders to gauge alternative future land uses thus making planning and decisionmaking
processes potentially more transparent and democratic. In this paper, we detail a new method
that enhances GIS-web-based public participation. We build on a combination of GIS basic capabilities
and the data mining methods of Artificial Neural Networks (ANN), namely Multilayer Perceptron (MLP)
packaged in a friendly (GUI) user interface that runs on the Google Earth platform. Users will be able to
articulate different spatial development scenarios for a specific area, to conduct sensitivity analyses for various competing scenarios and to explore causal connections between them
To implement a GIS-ANN-MLP model that runs through a Web application we opted for a free open source neural network library, named Fast Artificial Neural Network (FANN). Before implementation, FANN was tested through the FANNTool GUI This enables us to prepare the data in FANN library standard, design, train, test and run the ANN. This enables us to prepare the data in FANN library standard, design, train, test and run the ANN.
