On the alternatives to smart cities, test-bed urbanism, and the possibility of replacing conventional statistics with big data.
The concept of smart cities has recently become popular worldwide: there is a growing number of city ratings; public authorities are deploying new technologies and competing to be “the smartest”; and the market is growing dramatically. At the same time, the general public is trying to understand whether these technologies make their life better.
Professor Rob Kitchin from the National University of Ireland Maynooth is one the most influential smart cities experts. With over 20 years of research experience, he has written or edited 23 academic books and 160 articles. He is a principal investigator on the Programmable City project and the Dublin Dashboard.
Professor Kitchin has two big projects at the moment, each of which is allotted several million euros. One is from the European Research Council and proposes to look at the social, political and economic implications of creating smart cities. The second one is Building City Dashboards, which just started in September. The idea is to extend the whole Dublin Dashboard project, shifting everything into open source so there will be no proprietary software involved in it and all the code will be available for any other city.
Anna Lvova, a Strelka Institute alumna currently working at Habidatum, discussed with Rob Kitchin whether citizens are involved in the decision-making process, how the data ontology problem should be solved and whether data repurposing has done more harm than good.
— The concept of the smart city is quite complicated. How would you explain it to your grandmother?
— I would say that it means “using digital technology to help manage the governance of the city more efficiently and sustainably.” In a sentence.
— When was the term “smart city” coined?
— We’ve been using digital technologies in cities at least since the 1950s within planning, and then GIS (geographic information systems) appeared; we had control rooms in cities for a really long time through SCADA systems (Supervisory Control And Data Acquisition - ed).
So, while we’ve got this label “smart cities” we’ve actually had this stuff going on before. If we go back into the 1980s people were talking about “wired cities,” then they were talking about “cyber cities” and then they were talking about “intelligent and knowledge cities.” People like Simon Marvin and Stephen Graham were writing about networked urbanism in the 1990s and the early 2000s.
The earliest you probably find the term “smart city” is around 2000, but it really entered the main lexicon in 2009-2010, which is when IBM started to use it as a part of their advertising campaign and when they were trying to promote new products on that basis.
— As soon as this concept entered public discourse, there also appeared wider criticism of it. What are the main concerns regarding the technology?
— It has been in public discourse for the last ten years. And people started to look at, well, what’s actually happening here:is this about new, very top-down technocratic forms of governance? Is it about the corporatization of governance? Is it about companies running parts of city infrastructure as opposed to cities running them? So, you kind of outsource or you privatize city services? People were asking: “Where is the citizen in all of this stuff? Who is smart city technology empowering? Is it just empowering the empowered or is it a democratizing technology?”
— What do you think?
— I think that the approach is pretty top-down and there’s very little citizen engagement. And where there is, it tends to be around things like hackathons or some public apps allowing people to report city problems and so on. The cities are not really involving people in discussions about whether we should deploy a new intelligent transport system, a new smart lighting system or whatever it might be. They tend to make these decisions themselves. It depends on the governance within the city, who’s making these decisions.
In Dublin a lot of decisions have been made by the city managers, who are bureaucrats. In places like Moscow, Boston, New York or Baltimore they’re made by a mayor. So it’s a political decision. And I think in both cases there’s actually been relatively little discussion with citizens. The technology just basically appears and then people can either accept it or protest against it.
— When I was looking through the program of the upcoming Smart City Expo World Congress in Barcelona I saw an interesting term: PPPP. Everyone knows about public-private-partnerships, but the organizers added the fourth “P”: people. Now that has to be articulated because it is not obvious that everything is made for people.
— You don’t want citizens involved in every single decision in the city, because it just becomes unworkable and they also don’t necessarily have the domain knowledge or the specialization to know whether something is a good technology or a good implementation. You do have to accept that you can’t have direct democracy on absolutely everything: that’s why we vote for politicians and appoint city managers.
At the same time there’s a long history of participation in the planning process. So, if you’re going to have a city plan, you’d normally go through a series of town hall meetings and consultations about what’s going to happen and so on. It’s still happening with things like land zoning and so on. But it’s not happening with these new technologies.
The technology just basically appears and then people can either accept it or protest against it
And people like Anthony Townsend and Adam Greenfield and Simon Marvin criticise the lack of citizen engagement, lack of transparency, corporate involvement and the privatization going on. It’s kind of a reinforcement of neoliberalism as opposed to using the technology to imagine and create a new kind of city: a more inclusive or engaged or participatory or emancipatory kind of city. I don’t think they’re necessarily against the idea of networked urbanism. What they are against is the way in which it’s being deployed at the moment.
— There was the influential idea of the “non-plan” suggested by the British architect Cedric Price in the ‘60s. It was a response to modernism, sort of a manifesto. Currently there is a lot of criticism of smart cities. But what is the alternative? Is there some sort of manifesto for non-smart cities?
— Do they have an alternative? That’s a good question. That’s actually a part of my lecture in Moscow. I also organized a workshop in September. I was trying to press people to say: “Okay, fine, that’s a problem. What would you do instead?”. And they couldn’t really give a lot of concrete suggestions. The third way I suggest is to be really clear about different things. Rather than saying: “This is the truth of cities, this is the city as it actually is,” it’s like: “This is one version of the city based on the data that we have and this data has these shortcomings.”
One of the problems with smart technology is that it’s very reductionist. It takes very complex relations and simplifies them, whereas actually we know that cities are really complex multilevel intertwined systems. For example, in our work with the Dublin Dashboards we’re saying there are issues with dashboards but we’re also saying: “This is what we think we can do about that.” Because they still have utility, so the danger is that you throw the baby out with the bathwater. It’s kind of like the traffic in Moscow. If there wasn’t an intelligent transport system, it would be 50 times worse.
— It is often said that cities vary widely from each other. Why exactly does that matter? Can you provide an example of a technology that works for one city but does not work for another?
— Open data is a good example. When you create your open data repository,almost every one has to be tailored to the city, because they’re using different data standards, formats and classification schemes. For example, there’s the 311 service in Boston: you can call and say the streetlight is not working or there’s vandalism and so on. Boston has 101 municipalities and each one of those municipalities has a different ontology for 311 data, which means you actually can’t combine that data together very easily. So there are problems even within the city.In Dublin we have four local authorities and each one has its own planning department and each one has its own land use classification scheme. So, pulling the four together to create a single land use classification scheme for the city is actually quite difficult, because they classify the land differently.
Standards are a huge issue everywhere. You can’t just take an Irish train and put it on a Russian railway because your tracks are much wider
We have the same problem with transport and mobility. One of the problems in Dublin with the Leap card (electronic ticketing used on public transport,like the Oyster card in London or Troika card in Moscow) is you actually have different service providers: bus companies, train companies, tram companies. And reconciling the payments across these different systems was actually very difficult. This is the data ontology problem. But it is really hard to change it. Whoever changes the data ontology loses the back series, the continuity. So, you might have a solution that works in Copenhagen but it doesn’t mean you can easily implement it in Dublin or Moscow.
— Should there be some international standards? Who should set them?
— Standards are a huge issue everywhere. I couldn’t take an Irish train and put it on a Russian railway. I think your tracks are much wider than our tracks. There’s a whole series of organizations like the International Standards Organization, the Smart Cities Council , and the British Standards Institute; they’re all creating smart city standards and there’s about 70 or 80 of these initiatives: I have a PhD student who is looking at these (Jim Merricks White - ed). There’s a new ISO standard for city indicators, for example; it was published in2014 and was meant to create a common set of indicators across cities across the globe. The interesting thing about that project when it started was that they tried to get indicator data from as many cities as they could and what they found is across all of the cities ‒ I think they did about 280 cities ‒ there were only three indicators that they could get for every single city. Only three.
Well, I’ve spoken to quite a few companies, and they said: “Can you actually come and tell us how cities work? Because we don’t actually understand how the city is organized”
— Seems like we need a standard for standardisation. How can we break the vicious cycle?
— I think it’s just going to be the way that standards have always been done in the past. At some point one company has such a dominance that they effectively set the standards and everybody has to follow it. You see this in the US and, I guess, elsewhere in the world, where they’ve overcome the data ontology problem on transport data. Because if you want your transport data to go into Google, you have to provide it in a Google format. And so, the company is so large, it goes across so many places, that they’ve actually forced all the cities to go onto their standard.
— What do you think about the growing influence and involvement of private companies in urban “smartization”? Do they always suggest efficient solutions?
— I’ve spoken to quite a few companies and the companies have asked me to go and talk to them; they said: “Can you actually come and tell us how cities work? Because we don’t actually understand how the city is organized politically and governance-wise and how they’re making decisions about what technologies to procure and not to procure and so on.”
For example, some companies have an Internet of Things solution like a nice sensor that will tell you the sound levels in the city. If there’s no noise directive that compels the city to do that then it’s almost a luxury. Dublin has a mandate to do homeless services. So if there’s a choice between homeless services and a sound sensor that’s optional then it’ll always pick homeless services. And some companies don’t necessarily understand that. And that’s why some cities have found the smart cities market more difficult than they anticipated. Cities have not been buying their services because they don’t necessarily see what the value-added of the technology is.
— In addition, these solutions are usually extremely expensive.
— Yeah, and there are also things like legacy infrastructure. Dublin was looking at smart lighting, so we did the scoping study. It’s very simple to just convert to LED, but you can’t easily layer the commander control system onto the electric system, because the cabling is 70 years old.
Cities can understand the LED light ‒ that’s really simple: take out one bulb, put in an LED bulb, you save 65% on your energy bill. Why do I want to be able to turn the lights up and down and dim and trim them and change the color, add a commander control system onto it? Do I really want to replace ‒ in Dublin’s case ‒ 47000 lamp posts with smart lightning? It’s going to cost me 15 million euros. Maybe I just want to change to the LED bulb?
— To continue with data corporatisation. There are a lot of experiments by private companies like Uber, for example, in a city in Florida, just trying to reimagine the whole city transportation. What’s your attitude to these private experiments within cities?
— Well, it’s a big area now, that notion of test-bed urbanism. Actually it’s one of the things that Dublin is trying to specialize in offering companies: the ability to test their technology in the city. On one level, I think it’s fine in terms of trying to produce better products. On another level, I’d be worried about the corporatization of key city data sets, so the only way you can get the data is basically to buy it.
In the recent period from around 2009, cities were a little bit bamboozled by the smart thing, they were a little bit caught in the headlights of all this new shiny technology. In Dublin when they did the bike share scheme, they didn’t think about the data, so when the system was implemented and they asked the company for the bike data, the company just said: “No, it wasn’t on the initial contract, we’re not giving it to you.” They lost control of the data accidentally cause they hadn’t thought about it.
Now I think the pendulum has swung back. The people and the authorities have started to understand that they actually have the key thing here: they have the city. And it’s them that’s procuring the services and, even for things like Uber, it’s them that’s in charge of regulating how the service is doing its work.
— Urban analysis deals a lot with data repurposing. Personally, I work with social media and studying cities through the lens of social media. However, there is a professional debate about whether it is ethical to repurpose data. Some people suggest minimizing the data usage: one type of data for one purpose. But it’s difficult to minimize the ways in which data might be used, because we are always extracting new meaning from data.
— Data minimization is really about protecting people’s privacy and issues like predictive profiling and so on. So, it’s about the data not being used in a way that they didn’t anticipate or being used in a way against them that they didn’t anticipate. When people are using Twitter, they’re not thinking that that data might be used in predictive policing or in national security. Data minimization is about using the data for the purpose for which it was intended. Now, for big data it’s a problem because big data is all about repurposing data for secondary uses.
— For example, we use open data to study human mobility, we measure the agglomerations by using social media...
— Exactly. It’s useful as long as it’s done in a way that’s sensible, so that there’s a proper de-identification and anonymization and maybe aggregation. The other thing is to be very clear that big data is still sampled unrepresentative data. So not everybody is on Twitter, not everybody is using the Troika card, not everybody is driving and has a number plate to be on automatic number plate recognition. It’s always still sampled data.
— But the same works with polls, surveys etc.
— Yeah, exactly. Other kinds of data are not necessarily representative, either. I have another project actually, which is about big data and official statistics and about the question of could you replace conventional statistics with new ones produced from big data? And one of the key issues is about the representativeness of the data. I know one European country that’s looking at Twitter data as a potential way to measure health and well-being. Well, the key thing to remember is only 4% of people over the age of 65 are on Twitter and people over the age of 65 are the people who are most in need of health and well-being issues. So, the key population is actually missing out on the instrument that they’re trying to repurpose. That’s a problem. And it’s about understanding that and compensating for that. But at the same time, if you’re trying to optimize traffic, you only actually do need to know people who’ve got number plates.