My colleagues Zhao Yiting and Long Ying from the Beijing City Lab have graciously translated our paper called Smart Cities of the Future. I owe them many thanks. It is available in a special issue of the journal Urban Planning International (1673-9493, 2014,  06-0012-19) and you can download both the original English version which is open access and the new Chinese version by clicking on ‘English’ or ‘Chinese’ in this text. Here is the abstract.

摘要:本文初步概括了智慧城市的组成要素。智慧城市是指运用新数字技术进行协同与整合,将现代信息通信技术与城市传统基础设施有机结合起 来的城市。首先,我们提出了智慧城市的7 个目标:(1)发现理解城市问题的新视角;(2)高效灵活地整合城市技术;(3)不同尺度城市时空数据 的模型与方法;(4)开发通信与传媒新技术;(5)开发城市管理与组织新模式;(6)定义与城市、交通、能源等相关的重大问题;(7)识别智慧城 市中的风险、不确定性及灾害。为实现以上目标,我们的研究需要在六个方面有所突破:(1)通过管理、控制和优化,使智慧城市的基础设施与实 际运行、前期规划更好的衔接;(2)探索城市作为创新实验室的新理念;(3)提供城市模拟技术目录,为未来设计提供指引;(4)探索更加公平合 理的技术方法以实现更好的城市生活品质;(5)探索促进有效公众参与以及公众认知的民主化城市管理新技术;(6)保障更加便捷高效的人口流动 性以及机会获取渠道。文章首先梳理了当前城市技术发展概况,并对智慧城市科学进行定义。我们将目前的智慧城市归纳为6 种情景类型:(1)旧 城的智慧型更新;(2)科技园建设;(3)围绕高新技术的科技城建设;(4)运用当前信息通信技术的城市公共服务;(5)运用信息通信技术开发新 的城市智慧功能;(6)运用网络以及移动客户端开发公众参与新模式。接下来,我们提出了七类可探索的项目领域:(1)智慧城市综合数据库的建立; (2)数据采集、网络分析技术以及新社交媒体的影响;(3)网络及移动行为建模;(4)城市土地使用、交通、经济互动的建模;(5)城市劳动市场 与住房市场交易活动的建模;(6)智慧城市的决策支持,如城市智能技术,公众参与式城市管理以及(7)智慧城市规划架构等。最后,我们期望通 过这项研究,转变传统的城市研究范式,并进一步探索促进智慧城市科学形成和发展的关键要素。


Visualising Ranks and Size in Space and Time


Alluvial diagrams were first proposed to represent changes in network structure over time. Robin Edwards from CASA has implemented the tool and has several examples from social and political arrays which he shows in his blog GeoTheory. Rosvall and Bergstrom’s popularisation of the technique in their paper Mapping Change in Large Networks  was published in PLoS ONE, volume 5, in 2010. The technique goes back a long way to many visualizations of flow systems where the flows change in size with respect to one another over time, and the so-called Sankey diagram is a variant of this to represent branching flows widely employed in energy studies.

Here we show how the technique can be used to show how a system of cities where each city which is located in a different location, changes with respect to its size over time. What we do is rank the cities from the largest to the smallest plotting the biggest city by its size on the vertical axis at the first time period, the second by its size above the first and so on accumulating the sizes until the entire population of cities has been plotted. If we do this at the next time period, some of these cities will have changed in rank and most likely they will have all changed in size. If the whole population of these cities gets larger, the graph will increase in terms of its accumulated population. Otherwise it will reduce. So in a sense it is like a layer cake with the largest city at the bottom and the smallest at the top. The first diagram on the left above is this plot, while the second on the right is the same but normalised with respect to relative – proportionate – change.

However what makes this tool so interesting is that the rank of these cities may change. If the largest city at the first time period becomes the next largest at the next time period, its position on the graph will change, and in this way we will see the cities changing in rank as well as size through time. In our UK data set of cities as primary urban areas, if we consider the largest city in the urban system which is London, this remains the top rank over the last 100 years but Manchester is the second until 1941 when Birmingham takes over. We show this section of the graph which is the bottom three streams of the graph shown above as follows.


What we do to make the diagram flow, is present these city sizes as temporal streams whose width is city size. These streams of course can overlap between time periods and this represents changes in rank. We also separate every stream from each other by thin white band and this makes it easier to read the flow graph. We call these graphs ‘alluvial’ because they bear a resemblance to alluvial fans that result from deposits in streaming water.

In fact we can use this technique to examine any m x n table which represents quantities that vary continuously over their m and n dimensions. These quantities may be discrete – that is, the continuum of the dimension may be divided into discrete values but as long as one of the dimensions flows in terms of the way its variable changes, we can use this to represent the stream. If both flow, we can plot the alluvial diagram in either direction and this makes the tool quite generic. In fact we use it here to visualize changes in city size through time. As we show illustrate it, it is probably as good if not better than the rank-clock that I introduced some 8 years ago (Batty, 2006) where one simply plotted the rank, not the size, as a traceline or trajectory around a clock where the clock goes from the start time t at midnight/noon around to the last time t+T at the next midnight/noon.

There are many embellishments we can make to this system. Robin Edwards produced this plot using R scripting and he is currently producing a very interesting online interface that can accept any m x n matrix of quantities and visualize them in the form shown above. One can manipulate the plot in terms of the way it is illustrated – the top rank can be at the bottom with the bottom rank at the top and vice versa. One can start at any time period and colour the streams differently and so on. When it is ready I will post another comment on how you can get access to the program which will be through a web page.

At the Crossroads of Urban Growth


A spectre is haunting urban growth relating to how development is financed. My current editorial in Environment and Planning B (December 2014, issue 6) discusses the increasing disconnect between demand and supply of new buildings using the example of Wuhan in China where the dash for growth is turning traditional notions of how cities get built upside down. The prospects are frightening for there are many parts of the world where the economies might crash as the link between demand and supply is broken. We may well be in line for a series of linked credit crunches, one of which follows another as government debt, inflation, and the printing of money are increasingly out of step with what is feasible and sustainable.

The conventional wisdom is that developers borrow cash to construct buildings and infrastructure for which there is demonstrable demand, at least in the medium term. Banks that are risk averse will only lend money to those developers who have not only good track records but a convincing argument that their products will find a market. This is little different from any production which is predicated, to an extent, on present and future consumption. By and large for the last 200 years, this process had led to balanced supply with demand, notwithstanding more than normal profits being generated quite frequently. But in many parts of the world, particularly where urban growth is very rapid, a disconnect between the rate of growth and development and the rationale for its financing has emerged. In fact the credit crunch in the west and globally began with financial institutions lending to all and sundry without any of the collateral needed to justify this lending. This is being perpetuated particularly in China, and my editorial discusses some of the consequences.