Data-Driven Urban Dynamics at ORNL

Visited the Urban Dynamics Institute (UDI) headed up Budhendra Bhaduri (Budhu) which is a rapidly growing effort in data-driven technologies applicable to cities. This group which is based at Oakridge National Labs which began life in the war years as part of the Manhattan Project, focusses on research that is geared to understanding, predicting and resolving key urban problems using large data sets which are spatially extensive. It is one of the few groups around the world doing serious research into the science of cities, albeit data-driven in that the focus is on extracting urban patterns and processes from very large data sets. Traditionally these data sets have been based on imaging from remotely sensed data but are now being extended to social media and all kinds of real-time sensed data dealing with location, mobility, transportation and climate. Many of these projects utilise the powerful computer technologies established at Oak Ridge by the Department of Energy that runs the National Labs. At Oak Ridge, the UDI used the TITAN supercomputer for processing remotely sensed images that require various forms of deep learning in the extraction of pattern. TITAN is the fourth largest supercomputer worldwide (measured by petaflops – floating point operations per second which is greater than 10pb) but it is being replaced by SUMMIT and the lab could well take the top spot, overtaking China again – who knows.

The UDI’s most visible project is Landscan. This is a world-wide data set at very high resolutions – at various grains down to approximately 1 km resolution (30″ X 30″). let me quote from the web page: ” LandScan is the finest resolution global population distribution data available and represents an ambient population (average over 24 hours). The LandScan algorithm, an R&D 100 Award Winner, uses spatial data and imagery analysis technologies and a multi-variable dasymetric modeling approach to disaggregate census counts within an administrative boundary. Since no single population distribution model can account for the differences in spatial data availability, quality, scale, and accuracy as well as the differences in cultural settlement practices, LandScan population distribution models are tailored to match the data conditions and geographical nature of each individual country and region.”

I am a member of the Scientific Advisory Board [myself, Mike Goodchild (UCSB), John Harbor (Purdue), Nigel Davis (Willis),  Karen Seto (Yale), and Luc Vincent (Lyft) were in attendance] and we visited Oak Ridge last week March 27-29 and I gave a talk to the Institute. Bhudu asked me to talk on Cities and AI and although I complained mildly that I didnt know much about AI, I have explored how neural nets can be used in models of design so I decided to talk about this. You can access and read my talk by clicking on this link that will give you the PDF

However the icing on the cake was a visit to see TITAN and the photographs above show a panorama of me along the bank of boxes which are thence arrayed in parallel in 8 rows, a truly massive machine. Also saw one like this at Harwell (Rutherford-Appleton) the other week at STFC !

 

In The Post-Urban World

This new book edited by Tigran Haas and Hans Westlund from KTH is a collection of interesting and somewhat oblique essays on the urban world we have entered. Lot of people you know writing here. Ed Glaeser, Richard Florida, Patrick Adler, Rahul Mehrotra, Felipe Vera, myself, Hans Westlund, Paul Knox, and Richard Sennett – and that is the first part. And then in two more parts: Jessie Poon, Wei Yin, Kaisa Snellman, Jennifer Silva, Carl Frederick, Robert Putnam, Kyle Farrell, Tigran Haas, Fulong Wu, Karima Kourtit, Peter Nijkamp, Edward Soja, Fran Tonkiss, Laura Burkhalter, Manuel Castells, Saskia Sassen, Susan Fainstein, Emily Talen, Michael Neuman, Nadia Nur, Nina-Marie Lister, Duncan McLaren and Julian Agyeman. You can get a sneak preview using some Google Gizmo that is attached to the site.

In the last few decades, many global cities and towns have experienced unprecedented economic, social, and spatial structural change. Today, we find ourselves at the juncture between entering a post-urban and a post-political world, both presenting new challenges to our metropolitan regions, municipalities, and cities. Many megacities, declining regions and towns are experiencing an increase in the number of complex problems regarding internal relationships, governance, and external connections. In particular, a growing disparity exists between citizens that are socially excluded within declining physical and economic realms and those situated in thriving geographic areas. This book conveys how forces of structural change shape the urban landscape.

In The Post-Urban World is divided into three main sections: Spatial Transformations and the New Geography of Cities and Regions; Urbanization, Knowledge Economies, and Social Structuration; and New Cultures in a Post-Political and Post-Resilient World. One important subject covered in this book, in addition to the spatial and economic forces that shape our regions, cities, and neighbourhoods, is the social, cultural, ecological, and psychological aspects which are also critically involved. Additionally, the urban transformation occurring throughout cities is thoroughly discussed. Written by today’s leading experts in urban studies, this book discusses subjects from different theoretical standpoints, as well as various methodological approaches and perspectives; this is alongside the challenges and new solutions for cities and regions in an interconnected world of global economies.

More Big Data and Cities

Two new volumes in big data, cities and regions, with a strong spatial focus. The first Seeing Cities Through Big Data: Research, Methods and Applications in Urban Informatics is edited by Piyushimita (Vonu) Thakuriah, Nebiyou Tilahun, and Moira Zellner and published by Springer. This has some 30 key articles on a variety of methods including one by Greg Erhardt, Oliver Lock, Elsa Arcaute and myself called ‘A Big Data Mashing Tool for Measuring Transit System Performance’.

The second book Big Data for Regional Science edited by Laurie Schintler and Zhenhua Chen focuses the argument on regions as well in particular regional science, and contains some 28 chapters including a foreword by myself. There is a lot of food for thought here and the two books provide rather a good snapshot of the state of the art.