Smart cities are beginning to emerge because technology has made it feasible to collect, store and analyze data on virtually every aspect of city operations. In particular, the availability of low-cost sensors, ubiquitous connectivity and affordable cloud computing means that cities can connect every street lamp, bus and municipal building to the Internet of Things. For example, the cost of cloud storage on Amazon Web Services has dropped from 15 cents per gigabyte per month when the service launched in 2006 to 2 cents per gigabyte per month today. And the Raspberry Pi Zero — a barebones 1 gigahertz Linux computer used for embedded computing — sells for $5, making it possible to add on-board computing to virtually any device.
Like other technologies, smart cities will evolve and mature over time. The earliest will provide basic insights from data and enable local leaders to engage in evidence-based governance. These efforts will be important, but they will represent only incremental change from what cities have already been doing. For example, Baltimore created its CitiStat program in 1999 to measure all municipal functions and improve oversight and accountability of city agencies. Early smart cities will have substantially more data at their disposal, but they will not necessarily use this data in fundamentally new ways.
The second stage of smart cities will use predictive analytics to identify patterns and forecast trends. These types of insights will be especially valuable to city planners and local officials responsible for improving municipal services and responding to changing demands. These cities will reduce downtime on critical municipal infrastructure by performing preventive maintenance on vehicles, bridges and buildings, and more quickly intervene when public health and safety issues arise. This stage will rely on powerful data-driven technologies, such as the systems that enable Netflix to offer movie recommendations and Amazon to suggest additional products for customers.
The third stage of smart cities will focus on using “prescriptive analytics” to use data to optimize processes automatically. Whereas the second stage of smart cities will be primarily about using data to supply insights about the future that will allow city leaders to evaluate different choices, this third stage will be about relying on algorithms to make many of these decisions independently. Much like a system of smart traffic signals uses real-time data to optimize traffic flow, these algorithms will help to automate more government functions and increase the productivity of municipal employees.
At all three stages of smart city development, there is an opportunity for city leaders to look beyond local needs and consider how they can design a smart city that will be part of a larger network of cities that share and learn from one another. On its own, a smart city can use data to track local trends, but as part of a network, a smart city can benchmark itself against a set of similar peers. For example, water and waste management departments can compare metrics to assess their relative performance and identify opportunities for change.
If they hope to successfully develop into learning cities, cities can begin the process of setting up to work jointly with their peers by participating in forums such as the Global City Teams Challenge, an initiative to bring together government and industry stakeholders working on common smart city problems. But longer-term change will require city leaders to reorient their planning to consider not only the needs of their city, but also how they fit into the larger network.
