Computer visualization is central to the social sciences where very large data sets are to be explored and where there is substantial emphasis on understanding using spatial-geographical patterns. It is crucial to simulation where it is essential to make sense of very large data sets quickly, where it is necessary to predict many alternative spatial patterns, and where the system is too complex to reduce all analysis to numbers. In this talk, we will demonstrate three rather different ways of understanding geographic patterns in space and time taken from projects that we are working on in CASA. We are developing traditional methods for urban simulation based on land use transport modelling but entirely through visual interfaces that enable us to visualise the inputs and outputs of the models geographically at any stage of their calibration and prediction. We illustrate this with our land use transport model for Greater London which we are developing to look at long term scenarios involving climate change as part of our cities work in the Tyndall Centre. We will then look at a rather different style of modelling showing how we can increase our understanding through iconic digital models based on 3D GIS and CAD, illustrating how we can assemble and interpret urban form through distributions of buildings using our Virtual London model. Lastly we will look at how we represent and simulate local movements of children travelling from home to school linking this to energy use and visualising this in a widely available user interface based on Google Maps. This sets the scene for the following talk by Richard Milton which introduces the GMapCreator and MapTube tools for widespread dissemination and visualisation of geographic phenomena.
|
