Alistair Ford has just returned from the fourth annual Applied Urban Modelling symposium (link) in Cambridge where cutting-edge research in spatial simulation of cities, including land-use transport interaction models, agent-based modelling, and big data analysis, was presented and discussed. The conference started four years ago as a reunion of the spatial interaction community of the 1970s to examine progress over the intervening decades, but has grown to include other types of dynamic urban simulation, visualisation, and data analytics.
Michael Wegener (formerly Professor of Spatial Planning at the University of Dortmund) gave a fascinating opening address looking at predictions made by models and modellers 20 years ago and how accurate they had been, This included a discussion of Aldous Huxley’s 1931 book ‘Brave New World’ and whether current advances in data-gathering technologies are taking us in the direction of the future described in that novel! Other presentations on the first day included an attempt to map Lima’s slums using UAVs (ReMap Lima), modelling of all 654 Chinese cities using a parcel-based vector cellular automata model, and a new analysis of Singapore using community detection from transport network flows using analysis of smartcard data (link). Light relief was provided by a presentation of the Imaginary Lines project, a cross between an art project and an infrastructure model! To end the first day Paul Waddell from Berkley presented the latest developments from the UrbanSim stable including GeoCanvas, (now available to download for Beta testing) which allows fast 3D visualisation of large urban datasets, and the developments of UrbanSim to include an IPython interface and PostGIS support.
Mike Batty from CASA at UCL started the second day with a review of progress on urban modelling and the big issues facing the community today (including a loss of faith in predictive ability, loss of parsimony, the need for money and information flows as well as people and goods flows, inclusion of data about transport trips for non-work purposes (which are only ~25% of journeys in cities now), the challenge of growing city regions and globalisation, and the rapid change in cities today). Paul Buchanan from Volterra gave a critique of transport economics, showing how wrong a lot of assumptions (i.e. Value of Time measures) are, before Coen Tuellings from Cambridge University presented his model of land economics and transport in The Netherlands. Pascal Perez from SMART in Australia showed his work on TransMob, an agent-based model of urban travel and residences. The day finished with a fascinating session on the redevelopment of the King’s Cross area of London, giving a real insight into the planning and development process and showing us all just how difficult it is to model the ways that decisions are made in cities!
The final day included sessions on standards and benchmarking (including ISO standards for urban data) and urban dynamics (with some network analysis of city street networks and their evolution). The day, and conference, concluded with a 2-hour round-table discussion about emerging issues, where it was agreed that urban modelling needs to focus on new applications, particularly the challenges posed by climate change and the need for sustainability. Challenges that we in the School of Civil Engineering and Geosciences are well-placed to meet!