Smart city projects start with sensors and platforms. Almost never with a specific street. That’s the problem.
In nearly every smart city project briefing, the conversation starts with technology: sensors, connectivity, platform integration, data governance. It almost never starts with a specific street.
That tells me something: The smart city, as a concept and as a unit of intervention, is pointed at the wrong scale.
The problem isn’t data. Cities have more mobility data now than they’ve ever had. The problem is that city-scale intelligence produces answers at a resolution that doesn’t reach the decision that actually matters: whether a particular person, on a particular street, decides to leave the car at home this morning.
That decision is made at neighbourhood scale. And that is where the most effective smart urban interventions have actually worked.
Generic perception of smart city: Futuristic city life with drones, holograms and driverless cars. Is it smart city?
City Data Answers the Wrong Questions
A real-time traffic management system tells you that congestion has reduced by 8% across the metropolitan area. It doesn’t tell the parent on a residential cut-through why this morning’s school run felt different, or whether it did at all.
The aggregation that makes city-scale data legible to a political audience strips out the resolution needed to change daily decisions. Data becomes a dashboard metric. The dashboard gets consulted at budget meetings, not at the moment someone decides how to get to work.
The Integration Gap
There is a second scale problem in smart city works that receives less attention. It runs along a boundary that sits inside every city block: the line between private land and public land.
Real estate developers are increasingly building intelligence into new residential developments. Energy management systems, occupancy sensors, parcel automation, EV charging networks — the technical infrastructure exists and the commercial logic is clear. Smart buildings sell. Developers who install connected systems can offer residents the kind of operational transparency that older housing stock cannot match.
But that intelligence stops at the property boundary. The street outside the building is managed by a public authority, with different data standards, different governance, and in many cases no connected infrastructure at all. A resident’s smart home system can tell them their energy consumption in real time. It cannot tell them whether the pedestrian signal at the end of the road is operating correctly, or when the green space in the next block was last assessed for safety, or whether the mobility data from their building’s bike store has ever been combined with the bus frequency data for their street.
The integration required is technically and legally complex. It demands shared data standards that most cities have not yet established. It requires legal frameworks to govern data sharing between private developers and public authorities. It needs governance models that balance resident privacy against neighbourhood-level insight. The commercial incentive structure that drives smart home development actively works against it: a developer’s default obligation ends at the plot boundary.
But the solution is clear: if requirements are written into the masterplan early enough, private smart home systems can connect to streets, green spaces, and public infrastructure. Without that condition, the intelligence stays locked inside the building. The neighbourhood stays fragmented. And the data that might change whether someone walks or drives to their destination in the morning sits unused, one property boundary away from where it would actually matter.
Bounded Scale Changes Behaviour
The Barcelona Superblock programme demonstrates what neighbourhood-scale intervention actually achieves. The concept is simple: close the through-roads in a 3×3 block grid to private traffic and reallocate the space to pedestrians, cycling, and community use. The “smart” element was the data that preceded the design decision: pedestrian counts, air quality readings, and resident surveys conducted at block level, rather than the infrastructure deployed to monitor the result.
What changed behaviour was the physical redesign of a bounded space. In the pilot superblocks, through-traffic reduced substantially and residents began using the reclaimed street space within weeks. Parents found routes to school they hadn’t considered before. The pavement became usable for people who had previously been avoiding it.
The programme worked because its feedback loops were tight. Residents experienced the change as immediate and complete within their block. Problems with the design were visible at street level and corrected quickly. Each intervention was small enough to iterate without disrupting the rest of the city.
London’s Mini-Holland scheme in Waltham Forest took the same approach. Rat-runs through residential streets were closed to through-traffic, and investment was concentrated in a defined area rather than spread thinly across a borough. Cycling rates in Waltham Forest increased significantly in the years after the scheme opened, and walking trips followed. The mechanism was identical: bounded geography, decisive intervention within the boundary, tight feedback from the street.
Neither programme relied on sophisticated city-wide sensing infrastructure. Both relied on rigorous data collection at street level, clear design decisions, and the political commitment to close roads.
The Helsinki Criterion
FORUM Virium Helsinki, the city’s innovation company, runs almost all of its smart city experimentation through Kalasatama, a regenerating district on the eastern waterfront being developed for up to 25,000 residents. The choice of scale is deliberate. At district level, interventions can be tested against a criterion that city-scale programmes struggle to maintain.
Not sensor coverage. Not platform integration. One question: does this give residents more free time in their day?
The programme’s stated target, an hour of free time per day for Kalasatama residents by reducing friction in daily services, is one of the most sophisticated frames for smart urban intervention. It refuses abstraction. You either get the hour back or you don’t. Every intervention, from mobility services to waste collection to household energy monitoring, is evaluated against the same human measure.
The Kalasatama model is exportable not because it uses particular tools. The tech stack changes as better options arrive. It’s exportable because it insists on the right unit of measurement. A city that asks “how many sensors do we have?” is asking the wrong question. A city that asks “do our residents have more time today than they did yesterday?” is on to something.
Smart Kalasatama in Helsinki. Tarrget: One more hour a day for all citizens, by giving efficient city services. Source: https://fiksukalasatama.fi/en/
Start at the Street
Smart city strategies have their place. Long-range network planning and city-wide emissions monitoring need city-scale thinking. I’m not arguing against that.
But if the goal is to change how people actually move through their daily lives, to reduce car dependency, to make walking and cycling the natural choice rather than the effortful one, then city scale is the wrong entry point. It’s too far from the decision.
Start with the neighbourhood. Name a specific block. Ask what changes when a resident opens their front door tomorrow morning and the space outside has been redesigned for people rather than vehicles. And ask whether the intelligence inside that building is connected — finally — to the street they are stepping onto.
That question is where smart urbanism lives. The rest is infrastructure in search of a reason to exist.