In The Multiplexed Metropolis, an article in the September 7 issue of The Economist, online business and finance editor Ludwig Siegele explores the impact of big data on cities. Cities are centers of social interactions, commerce, entertainment and many other human endeavors. Will the vast amounts of data generated by these activities, - properly collected, analyzed and acted upon, - lead to a kind of second electrification, transforming 21st century cities much as electricity did in the past?
“The power cables that penetrated cities in the late 19th century transformed their shape (there are no tall buildings without lifts), their transit systems, their nightlife, their sewerage (cities need a lot of pumps),” writes Siegele. “Ubiquitous data services might have impacts as wide-ranging: they could make cities more liveable, more efficient, more sustainable, perhaps more democratic.” But, he adds a note of caution: “Enthusiasts think that data services can change cities in this century as much as electricity did in the last one. They are a long way from proving their case.”
In the last few years, a growing number of initiatives have been launched in business, government and academia to help prove the case for data-driven smart cities. Last year, for example, New York City Mayor Michael Bloomberg announced the formation of NYU’s Center for Urban Science and Progress (CUSP), a partnership between NY City agencies, NYU and other academic institutions, and a number of companies including IBM, Microsoft, Xerox, Cisco and Siemens. A few weeks ago, mayor Bloomberg personally welcomed the inaugural class in CUSP’s MS in Applied Urban Science and Informatics program.
Urban informatics, succinctly defined as Big Cities + Big Data, is a complex new field requiring considerable research and experimentation. There will be a variety of approaches to urban informatics, much as there have been in urban planning. Over the centuries, we’ve seen a number of different models to help guide the orderly development and growth of urban environments. Urban planning emerged as a scholarly discipline about 100 years ago to help deal with the accelerated, and often chaotic, growth of cities during the Industrial Revolution.Among the various theories for urban design, two major approaches stand out: top-down, central planning; and bottom-up, grass-roots planning. In mid-20th century New York, these very different styles were personally exemplified by Robert Moses and Jane Jacobs respectively, and led to heated debates about their merits and pitfalls.
The debates continue. The Economist notes that: “The use of data in cities pits top-down against bottom-up in a similar way. One side stresses the need for citywide planning and control, the other advocates just providing access to data that lets citizens make their own decisions.”
These top-down versus bottom-up arguments are being played out across most disciplines involving highly complex sociotechnical systems, that is, complex systems which combine powerful, inexpensive and ubiquitous digital technologies with the people and organizations they are transforming. Sociotechnical systems generally exhibit a level of complexity that is often beyond our ability to understand and control. Not only do they have to deal with the complexities associated with large scale physical and digital infrastructures, but with the even more complex issues involved in human and organizational behaviors.
When it comes to data-driven sociotechnical systems, - not only as they apply to smart cities, but to engineering, health, government, business or economies, - are we better off embracing a top-down, more central approach, or a bottom-up, more distributed approach? What kinds of architectures, if any, are best suitable for such systems?
Top-down approaches have worked well for static or relatively slow changing, well understood systems. A few basic engineering principles have emerged over the years, including the hierarchical decomposition of the system into relatively independent, functional modules and components; and the use of well defined processes for the assembly, testing, operation and continuous improvement of the overall system. This is possible, - even with highly complicated systems like cars, airplanes, bridges and skyscrapers, - with systems that are deterministic in nature, that is, their future behaviors can be generally calculated. For the most part, they will produce the same outputs from a given set of inputs. No matter how complicated they are, the whole is pretty much equal to the sum of its parts.
However, such an approach breaks down for complex systems composed of many different kinds of components, intricate organizations and highly different structures, all highly interconnected and fast changing. Such systems exhibit dynamic, unpredictable behaviors as a result of the interactions of their various components. This is the case with cities and other complex sociotechnical systems, where people are the key components. People, communities and social organizations organizations exhibit far more variations than is generally the case with physical systems. With complex systems, the whole can be far different than the sum of its parts.
“Although many such [top-down] systems are supposed to work automatically, it is a rare smart-city project that does not aspire to a NASA-style control room filled with electronics, earnestness and a sense of the future,” writes Siegele. On the other hand, “From the bottom-up view, the control room is a smartphone. Devices that know where they are have allowed enthusiasts to build all kinds of new applications . . . These come into their own in the dense social worlds of cities. . .”
“But this enthusiasm has rarely, so far, translated into game-changing success: except in the area of public transport, few apps using open data have made the jump from interesting novelty to reliable consumer service. Venture capitalists have not proved very enthusiastic about them; many developers have given up. The data provided by cities may be free, but they are often poorly formatted or lacking in necessary metadata—such as details of location. Commercial data cost money.”
While the bottom-up approaches to complex systems generate lots of good ideas, they also have serious limitations. First of all, mission critical applications, including those used by the police, fire department and emergency management need to be very carefully designed and operated. These applications are intrinsically top-down. So are financial management applications in general, including tax collections and social assistance payments.
Privacy is another major consideration. Its important that government, at all levels, make more of their data openly available. It increases transparency and enables citizens to develop all kinds of innovative apps to complement the government’s own applications. However, given the enormous amounts of personal data out there, as well as the ability to link and find patterns across multiple data sets, we must be really careful that such open data, released with the best of intentions, does not lead to serious violations of people’s privacy.
While innovation almost always emerges bottom-up from individuals, research communities and start-ups, their successful implementation and deployment requires a fair degree of the kind of top-down governance best provided by companies, government agencies or professional organizations. Even open organizations like the Internet, World Wide Web and Linux have succeeded because they carefully balance the bottom-up innovation of their highly creative communities with the top-down governance provided by IETF, W3C and the Linux Foundation respectively.
There is general agreement that platforms are the most effective architectures for achieving such a balance. Institutions, - whether private, public or NGOs, - are typically responsible for the development, operation and governance of the platform. They make the key business and architectural decisions, including market segment and target audience, as well as standards and interfaces.
Platforms aim to attract a wide ecosystem of users and companies, which will in turn develop a wide variety of innovative applications and services on the platform. The more such applications and services, the more valuable the platform becomes. The bulk of the innovation will generally come from the ecosystem, while the governance and support comes from the platform providers.
Platforms will play a major role in the evolution of smart cities, notes Siegele’s article:
“Many of the people actually setting up information systems for cities see their job as providing a platform, ... City governments could provide basic services such as environmental and traffic information, a citywide payment system along the lines of those now often used for mass transit and firewalls to keep users safe from hackers and other digital mischief-makers - and let citizens and companies use them to build their own offerings. But the balance between what the city provides and what is sorted out by the citizens and the companies they do with business with will differ from place to place.”
Different cities will make widely different choices on how to leverage their data services, largely based on cultural and political factors, as has long been the case with urban planning in general.
“And that variety should, in itself, act as a safeguard against dystopia. One of the great things about cities is that they can and do compete with each other. In most countries people have at least some choice as to which city to live and do business in. The quality of the information platform a city offers will increasingly become a factor in those choices.”