Big data and machine learning are helping cities bring new services to citizens.
The value and volume of data has been increasing sharply in recent years, and over the decades to come will become be vital to a city’s operations. But handling, managing, and capitalizing on this data will be a challenge. This is where technologies like big data and machine learning come in.
SmartCitiesWorld recently spoke to Florian Von Walter, Solutions Engineering Director at Cloudera, and Douglas O’Flaherty, Global Ecosystem Leader at IBM Storage to talk about the value of data and which technologies will help cities take advantage of it.
Graeme Neill, editor, SmartCitiesWorld: Why are big data and machine learning so important to cities today?
Florian von Walter, solutions engineering director Cloudera: In the context of electricity, mobility, public transport, water supply—all the basic needs that a city has had to fulfil to serve the basic needs of its inhabitants - all of this data has been generated for years. It’s not getting less, it’s getting more because of the increase of new technologies that are being introduced, and because of the growth of the cities. Especially for large cities.
The more people you get in a city, the more data is being generated, and the more management and administration needs to be done. From that perspective, more and more data are being generated but it’s probably being generated and stored in different types of formats, in different organisations. It might be siloed in the sense that multiple organisations in the city exist and they’re sitting on the data but they’re not bringing it together to correlate it with each other to get more benefits out of the joint data. From that perspective, big data certainly is something that can benefit big cities, or all types of cities.
Coming to the second part of your question about why machine learning is so important: because of the increase of the data and because of the various data sources that are being ingested, it’s important to make sure that the data are being analysed properly, and machine learning can be an important part of this analysis process, and also getting new insights of the data that is being collected.
SCW: Douglas, why these technologies are important today, but also why they’re going to be increasingly important over the next decade?
Douglas O’Flaherty, global ecosystem leader, IBM Storage: When we think about this, the citizenry of the current city is changing as well. We’re expecting to see things that are much more interactive, we’ve changed our relationships with our governments, and we’re looking for optimization. As Florian referred to, cities are growing, they’re very vibrant today but that also means new challenges. How are we managing energy? How are we managing transport? How are we introducing new types of better living? One of the things that happens when you can bring together a large quantity of data is that you can look at it from different angles. You can really pay attention to what are the clusters or what are the values that you have there.
Unintended consequences can often come out of bringing together good ideas and taking a look at them in new ways. It could lead to insights so that people and the citizenry and the cities can actually operate more efficiently, more humanely, and more modern in a way. And that saves everybody money, it saves everybody time, and actually makes for a better government.
Cities are growing, they’re very vibrant today but that also means new challenges
IBM runs a process that we refer to as “design thinking” and we often bring in urban planners and groups together to think around the many ways that you could improve something. It’s really the data that helps validate and drive some of the insights that can be brought from that.
SCW: Why are these technologies going to become increasingly important over the next 10 years?
von Walter: Well, I think there are many examples that we can name, explaining why this kind of technology is getting more important. For example, a parking space is becoming a more and more scarce resource in cities because of the increase in traffic, the increase in the number of cars, and in general, the amount of transportation in the city. Whenever you’re looking for a parking space, it’s getting more and more difficult to get one. Introducing or using these types of technologies to advise people where the next parking space could be free is one example.
And that basically brings us to the next incarnation of how big data, machine learning, or technology in general, can play a role because connected cars today are a reality. It’s just a matter of how you connect them and what kind of information is being sent to the cars or exchanged between smart city infrastructure and the car itself. For example, an indication of where the next free parking lot could be via these connected capabilities in the car is another good example.
Another example could be monitoring air pollution in the city. I think that’s a pretty hot topic nowadays. Monitoring air pollution and then controlling how many cars could enter a city, for example, at a certain time of the day, could be a way to decrease air pollution and maintain proper mobility and transportation throughout the city.
SCW: How do you think cities are going to be using big data and machine learning?
O’Flaherty: I loved Florian’s example of parking because it’s a good example of how data structures and managing of the big data comes back to some of the reasons that IBM and Cloudera are working together. These are very big and dispersed data problems. When it comes down to other ways that we’re going to use it, clearly the social citizen, the single view of the citizen, is part of that environment as well. We’re trying to understand how flow and new construction can change the way the citizenry interacts with their own environment.
The parking lot example is a really good, easy way to look at some of the big data problems. If any of us have been into a modern parking lot, there are lights that will tell you if a space is open or not. In my hometown here in the US, I have an app that allows me to pay for parking from anywhere. So, to Florian’s earlier point, we’re collecting this data and ingesting it from many different areas, and each one of those is smart already. It knows what its local parking is, it knows what’s happening in the parking area, it knows where the data is being ingested to. But then we bring that together in more of a filtered way and put them all together into a larger view, and suddenly you have a connection between what was happening for tolling or what’s happening at a tunnel, what you’re seeing with air pollution growing in a certain area, and a lack of a parking space.
Urban planners could start to look at either a different parking garage, or perhaps my preference, which would be a rideshare, better trams, better trains, light rail, or a place to park bicycles. Those become part of the way that you can use big data and machine learning to pinpoint congestion points in multi-mode traffic that aren’t available to you through conventional models or traditional approaches to urban planning that tend to focus on single mode of travel. That’s one of the reasons why I thought that Florian’s example of parking was excellent. But you can take it into lots of other areas: water management, social education, and others as well.
SCW: Can you elaborate on those a little bit for me, please?
O’Flaherty: For example, in education, we are always—especially here in the United States, and we’re seeing this all over the world—we’re trying to understand what is the best way to educate our children and what are the environments that they’re in. What makes for a good environment, especially in early education? I have a neighbour who actually specializes in that. What she’s looking at is, neighbourhood-by-neighbourhood, what are the things that raise up a child in education? A proximity to libraries, families that read, certain programs that work. We can track that, and we do today. We have digitised huge portions of that process and then we can track that cohort through several years. By looking at what has worked and what has not, there’s insight into the citizenry that can come from that.
We see it less here in the United States but we see it all over in other countries where you see unified healthcare systems. For example, there’s some very good work that IBM is doing with the NHS in the UK, trying to understand, digitise, and track what’s going on. All those things help, perhaps not at the level of individuals but help at the level of cohorts, or help at the level of the citizenry.
Those are some of the other examples that I think are very under-tapped today, because often they feel like they’re underfunded. People think of this AI and machine learning as hard work, and it has been, but the frameworks that are in place—the things that Cloudera is bringing to market, the things that IBM brings together, the way that we can manipulate and move data, the ability to manage large quantities of data—make this much more accessible than it ever was before. And we’re raising an entire citizenry who is digitally savvy and has the ability to interact and therefore can provide us with more information about how they’re doing, and collect more information, who want to have that level of interaction with their governments.
People think of this AI and machine learning as hard work but the frameworks that are in place make them much more accessible than they ever were before
SCW: Florian, Douglas has raised a very interesting point about cities that see AI as hard work. If a city decides not to explore big data and machine learning and not put them at the heart of their smart city strategies, what are they going to be missing out on?
von Walter: Well, I guess that’s a little bit difficult to answer, because if you look at large cities, they cannot avoid looking into big data and machine learning, or exploiting new technologies that make living in these cities worthwhile, and improving living circumstances. Coming back to air pollution as an example, or other negative circumstances that might come up from the growth of the cities.
On the other hand, it’s really difficult to say that cities are missing out on something because a lot of the data already exists today. But big data and machine learning can just make it easier to deal with the consequences. If they’re not doing that, then these cities are probably becoming less attractive in terms of getting new citizens. It’s just a matter of what kind of future they’re envisioning for their cities.
For example, if you look at things like autonomous driving, autonomous buses, and autonomous infrastructure, it’s certainly important that big data and machine learning are playing a role in this. Because otherwise, these technologies won’t function properly, or you will have a hard time to actually make them work together and play together nicely with the existing infrastructure.
From that perspective, cities that aren’t looking into these technologies might miss out in the future in terms of attractiveness.
SCW: Douglas, would you agree that cities will be less attractive if they don’t pursue these technologies?
O’Flaherty: I do think that Florian is absolutely right that cities that don’t pursue this will find themselves less liveable, but there’s another point that Florian brought up that I thought I’d circle back to, which is: they’re already collecting this data. The data exists, they’re paying for it, they’re throwing it on—I’m a storage guy so I keep envisioning how much money they’re spending—they’re throwing it on slow old stuff that they’re never looking at. They’re unlocking the data that they already have.
There’s a significant step between a whole pile of smart meters and some feedback on what traffic patterns are looking like to building a whole data lake environment that you can work with but the data is there and it’s already ready to go. They’re spending money on something without getting value from it.
There’s the first problem, which is upgrade your environment to the way you’re looking at it, to look at this as an asset, not a cost. That has been one of the biggest transformations with AI and ML. It’s the ability to look at that dark data, things that are currently sitting in places, and bring them together into a scalable infrastructure that you can look at and work with a small-focus team to really get some insights from.
The other part that Florian also hit upon, about liveability, is that the city becomes less adaptive. What happens when…? What if there’s…? How do you plan for a new sub-division? A new tramline? What happens when a water main breaks on a significant toll road, and how could that change the quality of life? And the ability for a city to feel like it’s always vibrant? Here we are in a pandemic doing this over Zoom, our cities are less vibrant but this is global. This all happens neighbourhood-by-neighbourhood as well, in terms of interruptions to daily life and cities not being able to adapt to it.
The data is there and it’s ready to go
One of the easiest examples is the prediction of buses. We put GPS trackers on buses today, there is data coming in to let us know, and there’s somebody who’s doing some basic analysis to see what’s going on. But with machine learning, and using an application with more compute power to it, you can start to predict certain routes that will become more congested as a result of something else happening in your city. You’ll be able to re-route, you’ll be able to accommodate, and as a result your citizenry will feel like it’s a more liveable place. I love the way Florian put it. Cities become more liveable when they can become more aware and more dynamic, and AI, machine learning, and big data can help them do that.
SCW: How do you predict big data and machine learning will be used in a decade’s time by successful cities that have it as part of their wider strategy?
von Walter: It’s very difficult to predict the future, but I hope that we will see that cities have started to use big data to increase the attractiveness of cities. The push to electric cars todays is a great example. Air pollution is being tackled by de-carbonization by moving to electric cars. Electric cars themselves can serve as power storage, which leads us to a discussion around how we can store electricity in the future in a better way. This leads us to the way we can utilize renewable energy generation, like as wind power or solar power, in a better way. Which again leads to the questions of how we can control power consumption in cities, and again, smart grids are a good example which is connected to that discussion.
All of these technologies need to be connected somehow to create better value and give us more effectiveness in terms of the resources that we are using. I hope to see that in about 10 years’ time, should we talk together again, that these kinds of technologies have been used or are in use to provide us a better means of living.
O’Flaherty: Well, it’s kind of fun to be talking about this here in 2021 because what we’re discussing is how a number of technologies have come together to be able to apply to an infrastructure that [already exists]. Smart meters and smart grids have been a project that has been underway for a decade or well over a decade.
In another 10 years, we will reach another technology inflection point as well. Florian referred to autonomous vehicles. What we’re going to see is a distribution of this data. Right now, we’re talking about bringing it altogether into a single space but when we start talking about autonomous vehicles, and areas, and neighbourhoods that can make their own decisions, we’re going to be talking about how the technology and the connectivity of things like 5G are really going to help drive forward the autonomy at the edge, and new human interactions that happen there.
In association with: