Skip to main content

How AI and advanced analytics can help save the world's cities

Cities are looking at how they can use emerging smart technologies and more efficient processes to make themselves more attractive places to live.

The world’s future lies in its cities. So notes a recent United Nations report, which projects that 68 percent of the world’s population will live in urban areas by 2050, up from 55 percent today. That’s an increase of 2.5 billion people in just over three decades. The report also warns that cities will face major difficulties in meeting the needs of their growing populations for housing, transportation, energy, infrastructure, employment, and basic services such as healthcare and education.

Given that so many cities today are clogged with pollution and cannot meet their residents’ most pressing needs, how can the world survive such a massive population shift?

The answer is clear: Cities have to turn to big data, cloud computing, advanced analytics, and artificial intelligence to solve their problems. That’s because cities don’t need to just get smart to survive the future. They need to become brilliant. 

The good news is that the transition to cities’ use of AI and vast amounts of data from the edge, IoT, and cloud is already underway. In this article, I explain how those techniques work and how they can yield operational insights to help tackle the challenging problems urban areas face.

It all starts with the data

The primary reason cities are getting smarter is dramatic advances in their ability to gather data. It’s becoming easier and cheaper for cities to install sensors, connect them, and send the data to a data center, where the data can be stored and mined for insights. Sensors will soon cost in the single digits and be powered by batteries with multiyear lifetime. They’ll communicate on 5G networks. They’ll be placed in many areas they couldn’t previously go. That means a lot more data from new data sources—even from, say, the inside of public garbage cans, enabling public works departments to monitor how full each is at any given time.

Compute power and storage are also becoming much cheaper, meaning cities can take advantage of all that new data and ask, “What can we learn from it? What insights can we gain from it?"

As for the AI side of things, a lot of the algorithms behind it have been around for quite some time—more than 40 years in some cases. But until now, we haven't had the cost-effective computational power, the large volumes of data, or the open source algorithms and models that let us run AI and machine learning. 

There are other influences driving the use of technology to enable better lives for city residents. For one, cities are starting to behave more like businesses as citizens become increasingly tech savvy, urbanized, and connected through smart devices and intelligent workplaces. Cities are examining how to be more appealing to citizens by treating them like customers and improving the quality of their lives. Cities are also trying to improve the efficiency of how they work. A lot of smaller cities, often called tier-two or tier-three cities, are looking at how they can start competing with bigger, more established tier-one cities by becoming more attractive places to live, leveraging emerging smart technologies and more efficient processes.

AI is a game changer. Explore what the journey involves, from building a roadmap through implementation and data migration.

Different cities, different platforms

Just as cities are different from one another, so are the platforms they use to become smarter. There’s no right or wrong answer about which platforms are best. But they all generally share a common basic architecture. As explained above, it starts with sensors deployed in the wild. Then there’s a communications channel that transmits the data from sensors to a data center. Inside the data center, there’s storage. Compute power interrogates the data and builds specific use cases from it. There’s also software, which can include everything from open source software to Python libraries to commercial off-the-shelf products. And finally, there’s a way to show the results of analyses so the data can be used.

In some instances, the data cities gather is made available to anyone. Madrid, for example, broadcasts its air quality data once an hour; anybody can download it and use the data for whatever they want. Some cities, such as Dubai, recognize their data has monetary value and charge people to access it. In such cases, the cities need to build a payment gateway and a data access gateway such as an API gateway or dataset downloads.

Using AI and data analytics to cut air pollution

How does all this work in practice? Let’s take a look at how a city can use AI and advanced analytics to tackle what has been one of the world’s most intractable problems: reducing the pollution caused by automobiles.

Cities need to be intelligent in how they manage traffic to cut pollution. They don’t want to just arbitrarily ban 50 percent of all cars coming in the city. Instead, they can use a combination of parking data, traffic data, air quality data, and weather data to make predictions about how cutting traffic will bring down pollutants to an acceptable level. Cities can then model different ways of reducing traffic to determine which would be most effective.

A fairly structured process makes that possible. First, descriptive analytics use historical data to examine air quality over time. Then diagnostics look for normal values and anomalies—in this case, what is normal air quality, what are the thresholds of higher and lower quality, and how is that associated with the weather, the number of cars in the city, and other data. That way, cities gain an understanding of what happens to air quality under various scenarios.

After that, predictive analytics powered by AI and machine learning can help predict what the impact on air quality will be by reducing the number of cars to a specific level. This is more complicated than you might imagine. For example, taking cars out of a city today doesn't mean it will see an instant drop in pollutants. There’s going to be a lag time before it starts to see the benefits in the air quality.

Weather has a big impact as well. For example, because Stuttgart, Germany, is set in a bowl, the weather has a major effect on how quickly pollution clears. If there's cloud cover, the pollution has nowhere to go. That means even if the number of cars coming into the city is dramatically reduced, pollution will still stay there.

Once cities build the models to predict what will happen in various scenarios, they can use machine learning to model solutions to the problems and predict how successful those might be. So, what’s the best way to restrict the number of cars coming into a city to curtail pollution? Charging congestion fees for people who drive into the city on certain days and times, like London does? Or handle it like Madrid, which alternates between allowing in cars with even-numbered plates and those with odd-numbered plates on heavy pollution days? The answer will be different for each city—and that’s where sensors, predictive analytics, AI, and machine learning come into play.

We’re still early in the process of using advanced analytics and AI to help solve pollution and other problems cities face. But it’s the way of the future—and the health and happiness of people across the world depend on it.

Related link: 

A smarter approach to smart cities: Unifying the use cases

This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.