“The smart city movement has been ongoing for a while, but it’s been more for the computer engineers. Most traditional urban planners haven’t really tapped into that huge resource yet,” said Mohamed Abdul Rahman of Hong Kong’s Polytechnic University, speaking to 6GWorldTM. A former urban planner himself, Abdul Rahman came to Hong Kong from Nigeria in 2015. “I wanted to see how I could incorporate Artificial Intelligence with urban planning.”
The main problem that he encountered? Data. “Surveying, the traditional method of data collection for our industry, is very time-intensive and very slow,” he explained. “For example, I had been collecting survey data for the past six months on a problem and had only been able to collect roughly 290 interviews.”
Much of his frustration stemmed from knowing there should be a better way to gather information on the quality of roads, the standards and availability of amenities, and access to shops.
“Instead of asking people what the problems of their cities are, almost all of these discussions have been held already online,” he said, “Why not just develop a system whereby people can easily download this information and be able to use it?”
The result was a project creating a system for accessing, refining, and validating urban social media data – published details are available here.
Data is also a challenge, in very different ways, for Sam Conrad Joyce, Head of the Meta Design Lab at Singapore University of Technology & Design. His team has been working with the Virtual Singapore project, which has generated a digital twin of the entire city-state. While this may seem like the apex of the urban planner’s work, it opens up very different challenges.
“At present, having a representation of the city is very much a first step,” he explained. “There are some questions that we can answer quite easily with this model, but answering many of the more interesting questions still takes a lot of additional work.
“For example, asking ‘If we put solar panels on all the tower roofs, how much energy could we generate?’ sounds simple. It would be, if all the roofs were horizontal, none of them shaded the others and the roofs had no other uses already. To get an aggregate for the city, you actually need to develop quite a bit of coding to account for all these factors.”
The Word on the Street
While urban planners have been conducting their own research, in many countries citizens have been finding ways to share information online about problems in their neighbourhood for years. For example, the UK witnessed a scandal in 2017: After Grenfell Tower in West London burnt down with the loss of 72 lives it came to light that residents’ groups had been raising concerns and blogging about the poor fire safety in the tower for more than five years.
Meanwhile, the Bey2ollak app was created in Cairo, Egypt in 2010 to help citizens navigate the terrible traffic congestion. Bey2ollak, meaning “the word on the street,” relies on users reporting traffic issues they experience to help others navigate around them. According to a 2017 interview with co-founder Gamal Sadek, Bey2ollak is a “community of positive commuters that took a conscious decision of helping each other out; we’re just providing the medium.”
With all this information out there, what has been holding urban planners back from using it? For Abdul Rahman, it’s a question of mindset as much as anything: “It was very hard for people in the profession to accept the [social media] data. Many of them think of social media as a lot of stuff and nonsense talking about celebrities or pets. In fact, if you look through it there’s a lot of sense hidden in there, even recommendations for how to solve problems.”
For Abdul Rahman, the way to change minds lies in finding the right ways to filter the enormous quantities of social media data to show its value. “The trick is to make access to this social media data very fast, very cheap, and to have the ability to do spatio-temporal analysis. For example, if I asked you when the last 10 times were that you went to a particular store you probably couldn’t answer. But I can go online, check your Google location data, and find out when you visited 100% accurately. So it’s more accurate than asking people.”
Conrad Joyce agreed that such data can be valuable, but not only that: “There’s definitely room for both types of information:top-down models and proposals as well as bottom-up discussion of pressures, problems, and needs. When you can match both together, that’s really the sweet spot.”
Soft Problems are the Hardest to Solve for Singpore
“Between different sources, we’re getting to a point where all the hard questions are answerable – that is, ‘hard’ in the sense that they have a specific and quantifiable answer, like ‘how long will it take to get from here to there?’ That’s pretty straightforward these days,” Conrad Joyce explained, but, for much urban planning, less tangible elements are in play.
“We know that there are mental health benefits from being around trees, but there are many other factors as well. Closeness to water; being in sunlight, but not full sun all the time; a breeze. While we have data on every tree in Singapore and can add or remove them from the model, we don’t know what impact they are having. Can you have a pleasant, restful experience in a plaza without trees? Yes, absolutely, but what’s the impact of adding them to the experience? Quantifying how these factors work together on emotional wellbeing would turn the data into something really smart and usable for urban planning.”
This kind of exploration is critical for the next stage in Singapore’s development, according to Conrad Joyce.
“What government figures are thinking about now is how we use the data to give us answers to softer questions. “Liveability” is a term much more in use these days. A place like Singapore doesn’t have a lot more space, so it’s hard to improve residents’ wellbeing by giving them bigger apartments or creating green space. Instead we have to look at other factors, like how to build a sense of community in an area. The next frontier is working out how to translate data into answers to that kind of question.”
Abdul Rahman is grappling with similar societal questions elsewhere.
“Currently I am using this on projects in university cities, because they tend to have more young students moving into poorer neighbourhoods. With time those areas gentrify and locals move out, and those younger people tend to be highly connected […] so I’m looking to see what types of challenge those cities are having from an urban-planning perspective. I’ve been able to look at sentiment data going back the past 10 years and get a sense of what challenges people have been having and what’s changed.”
His findings so far have been remarkably consistent across different countries: “Residents say the students don’t contribute, they’re not invested in the community. They just use resources, make a disturbance, leave litter, and move on. It undermines the happiness of permanent residents.”
From Urban-Planning Data to Solutions
When even a highly-connected smart-city leader like Singapore is grappling with the softer challenges of urban life, what are the next steps? Conrad Joyce is sanguine that, over time, the solutions that are evolving today will deliver real improvements.
“There’s a typical progression that people talk about, from data to information and then to insight,” he explained. “I think over time we’ll find a lot of data we’re collecting is actually pretty useless or useful only in very niche situations. We’ll get better at understanding what really is helpful and use that for real insights.”
Abdul Rahman agreed, adding that there’s a need to put the right tools in the right hands: “In most of the non-programming professions, such as my own urban-planning profession, most people don’t know how to code. We really need to teach people the easiest way to get out this kind of useful data from so much noise that’s being generated. So, what we did is create a library [of analysis tools] and put it on GitHub. With a minimum knowledge of coding people can use these tools to access data, clean it, and model it. Finally you can download it into [Microsoft] ExcelTM, because at the end your huge big data should become small data. That way you can move it around quickly and really find it useful, even if you don’t know coding at all.”
That said, Conrad Joyce urged caution: “Urban planners are limited, to an extent. We’re mainly dealing with infrastructure that has a long lifespan, so any changes are incremental. In addition, while a city can replace a building easily enough sometimes, if you have a protected building where you ideally need open space what can an urban planner do? Sometimes there’s a problem that just doesn’t have a ready solution.”
Will we see a change in the way urban planning works? “I really feel that the city can talk back to professionals,” Abdul Rahman commented. “I want to make sure I can develop this framework further so that, instead of just listening to what people are talking about, you could unite it with other data about a city and, for example, see in VR exactly where people are when they are talking about a particular problem.”
For Conrad Joyce it may come down to the telecoms industry. “Really, this is where the greatest stores of data are, updated every day by the citizens and the smartphones themselves. It’s an opportunity not just for data collection but, potentially, active participation from citizens. We can consult and crowdsource ideas. We just need to find a framework – ideally internationally – to make the data management and privacy work for everyone involved.”
