Then and Now

newnewthen

In November 1986 I visited SunYatSen University and gave a public lecture about Urban Modelling. China was a very different world then, no cars, no computers, no email, barely functioning electricity. And of course it was before laptops, networks, hand-held devices and so on. The personal computer had only just been invented. The campus was more or less in the countryside. Despite China opening up in 1979, this was still the old China.

Fast forward 31 years to 2017. The University is now a power house, in the top 10 in China and advancing in the QS university rankings worldwide very rapidly. Since 1986, I have been there a number of times but I never gave any more public lectures until last Tuesday and Wednesday when I more or less repeated what I had talked about 31 years ago. Well not quite, of course; it was the same domain of interest and in the same lineage – but I talked about web-based, large-scale urban models, ideas about big data, smart cities and so on. A world away from those distant years but closely linked intellectually.

newthennew

There are no faculty left still working in the School of Geography and Planning from those years but this is not unusual because there is no one left in any of the universities I have worked in before 1990. And it is a little sad that of those who were students then and now senior faculty there, none could remember attending my lecture and I am sure they did not know of it but there were about 130 in the room at the time. I found the building I had lectured in largely because of the poster above which was hand painted for my 1986 visit. Those of you who are Chinese will be able to read this.

Here are the pdfs of the presentations I gave:

Click on these and enjoy.

Modelling World 2016: Complexity in Land Use Transport Models

modelling-world

Giving a paper on Thursday 2nd June in London at Modelling World. Talking about Complexity in Land Use Transport Interaction (LUTI) Modelling, outlining very briefly our QUANT model for the Future Cities Catapult. This model is designed to simulate employment population and the interactions between and the constraints imposed on development for all of England and Wales at the scale of middle layer super output areas. There are 7201 such areas in E&W so to the scale is pretty typical of LUTI models. The model is web-based meaning that you can run it from anywhere and it is designed for anyone in E&W to explore the impact of changes in employment and population, networks time and costs and also land use constraints. Why only E&W? Well we are working in getting Scotland into it but the data is a little different when it comes to the journey to work and related census geographies.

It is early days as yet and we are still very much in the experimental stage of building this – it is in fact a proof of concept – to show we can build tools that are applicable to everywhere – and so far there are very few models like this one. We are stretching the state of the art in that the model, the data and the user(s) all interact with each other across servers and clients. The data and storage and speed implications of all this are pretty immense. More on this once we develop it further. I will post the talk after I have written it but before the session on June 2nd at 16-30pm at the Oval. Drill down for the content of the meeting etc. And here for my PDF of the talk which is hard to see in the meeting

Classifying Models: What Krugman Says

mathmodels

You may think that this blog has enough on ‘models’ to sink a ship and I have a new editorial on this in EPB. But there is a wonderful article by Paul Krugman on models which I have just come across although written two decades or so ago. Krugman compares evolutionary theory with evolutionary economics in a talk given to the European Association for Evolutionary Political Economy in 1995. The talk is republished here and although it says much on evolution and economics, he has some choice comments about how we should deal with models of the economy (and cities I hesitate to advance). In essence, he says that simulation models must always be interpreted from some meta model, some toy model that abstracts to the point where one really has two models: the full blown simulation and the handy reference – the toy model. He says: “By all means let us use simulation to push out the boundaries of our understanding; but just running a lot of simulations and seeing what happens is a frustrating and finally unproductive exercise unless you can somehow create a “model of the model” that lets you understand what is going on”.

In short he argues that we need models as metaphors and he concludes that: “In short, I believe that economics would be a more productive field if we learned something important from evolutionists: that models are metaphors, and that we should use them, not the other way around”. As all our theories are abstractions, then models represent the way we make these abstractions more practical, more operational. I have had another go at reclassifying models in our urban realm largely because they are continually morphing into new forms and the emphasis is changing, as much because the focus of our interest in cities is changing – witness the growth of the smart cities movement. So you can read the editorial here but Krugman’s essay is really the essential reading.