Dr. Philip Odonkor is an Assistant Professor at Stevens Institute of Technology’s Charles V. Schaefer, Jr. School of Engineering and Science, leading the Design Informatics Lab. He earned his PhD and MS in Mechanical Engineering from the University at Buffalo, SUNY. Dr. Odonkor’s research encompasses urban informatics, design optimization, cyber-physical systems, and sustainability. He received the prestigious NSF CAREER award in 2024 for advancing energy equity in urban areas. A co-founder of Grid Discovery, he is an active member of IEEE, ASME, and ACM. His work has been featured in Time Magazine and on the TEDx stage.
Odonkor discussed the intersection of informatics, smart cities, and sustainability. He emphasized that cities are complex socio-technical systems with inherent inefficiencies, particularly in energy use due to historically piecemeal infrastructure development. Odonkor detailed how data science and machine learning, especially reinforcement learning, can optimize energy consumption and improve urban living. He highlighted challenges like balancing privacy with data collection and integrating cultural values into algorithmic models. Odonkor noted varying global privacy attitudes and projected rapid electric and autonomous transportation advancements by 2025. He stressed the importance of long-term, adaptive urban planning.
Scott Douglas Jacobsen: Today, we’re here with Dr. Philip Odonkor, and we’ll discuss informatics, smart cities, and sustainability—a range of topics he specializes in. Let’s begin with the basics. What is the link between informatics and sustainability in cities?
Dr. Philip Odonkor: That’s an insightful question. Cities are inherently complex environments where people interact daily. While we typically notice the visible aspects of cities—residents, buildings, and transportation systems—examining cities closely reveals them as intricate socio-technical systems.
This viewpoint is central to my work. Complex systems often come with many inefficiencies. In most cities, you can observe inefficiencies in resource management, public services, and energy use. Cities have evolved over many decades, often adding infrastructure piecemeal to meet growing demands, leading to various inefficiencies.
As cities expand, energy systems are retrofitted to support the increased load, which can result in unsustainable practices and higher emissions. This is where informatics and data science—my focus areas—play a vital role. We can better understand how cities function and identify ways to improve efficiency using data. For instance, I analyze how buildings use electricity and seek methods to optimize energy use, ensuring that each unit of electricity contributes more effectively than it currently does. We can enhance sustainability and other critical urban metrics by addressing these inefficiencies.
Jacobsen: How do democratic systems impact long-term city planning? Democratic societies often operate within election cycles, limiting the ability to plan long-term projects. The focus on election terms can restrict the scope of planning, leading to infrastructure projects that may be segmented or delayed, resulting in inefficiencies as they progress through different political agendas. How do you consider this when redesigning energy systems and infrastructure?
Odonkor: I adopt a long-term perspective when seeking solutions in my work. Rather than focusing solely on short-term outcomes, I consider what we want cities to look like in 20 or 30 years. With this vision as a target, I work backward to determine the steps needed to bridge the gap between current systems and the desired future state. This approach helps guide decision-making and prioritize improvements. Although the process is complex and gradual, short-term strategies are integrated into this larger vision to make incremental progress toward a sustainable and efficient urban future.
There are two different scopes you can look at this from. I approach it from a long-term perspective, which helps smooth out some of the issues that might be apparent when looking at it from a short-term view.
Jacobsen: We live in an era of big data and systems that can process vast amounts of information. It would be beneficial to understand how to make sense of it all. Additionally, smart people build algorithms that can analyze and interpret this data. How do you gather data about a society or a city, and how do you make it understandable so you can use it effectively?
Odonkor: Those are great questions. How do I gather the data? Looking around your home or city, you’ll notice many devices capturing various metrics. For instance, some cameras, such as ring doorbells or surveillance cameras, record video. Across cities, we have sensors that capture data like temperature and noise, providing insights into city functions at any given time. This collection of sensors is part of the Internet of Things (IoT). IoT sensors are generally small, low-power devices that capture real-time data.
In recent years, there has been a significant increase in the number of these devices, making data collection much easier than before. However, the main challenge lies in handling this data. Just having data does not automatically lead to solutions. It often requires extracting insights from the data or combining multiple datasets to derive valuable information. This is where my research comes in—understanding what types of data to combine, when, and on what time scale to draw meaningful insights.
One way I approach this is through machine learning and artificial intelligence. I use machine learning extensively in my work, particularly a method known as reinforcement learning. Reinforcement learning allows us to program an algorithm to analyze data and understand the decision-making processes that generated it. For example, we have data about a home. In that case, the algorithm tries to determine what control actions led to certain patterns of electricity use.
Programming these algorithms aims to learn how a home uses energy. Once we achieve that, we can modify the algorithm to experiment with different actions and identify which changes could improve energy consumption within the home.
That way, I can start automating functions within your home so that it behaves as you would control it, but more efficiently. I understand what you prefer and don’t, and I can adjust things so everything appears normal. However, behind the scenes, processes are happening more efficiently.
Jacobsen: How do the niches within cities factor into this? For example, take Vancouver, which is near where I live. It has pockets of different subcultures, activities, and institutions, like universities scattered throughout the city. Do machine learning and AI algorithms naturally consider those variations part of their process?
Odonkor: No, not naturally. You can think of machine learning as an open canvas—you must tell it what to focus on, prioritize, and its objectives. Some machine-learning versions can attempt to figure things out independently. Still, in this domain, we guide the algorithms throughout the process.
It’s interesting that you bring this up because part of my work involves tuning algorithms to consider aspects beyond energy efficiency. For example, we design algorithms that balance efficiency while considering energy equity issues. Suppose we optimize energy usage in one area. Can the algorithm assess the demographics and decide where to prioritize energy distribution to balance overall usage? We also overlay this approach with considerations for access to renewable energy resources.
You can focus these algorithms on different topics. One of our key goals is to get these algorithms to incorporate energy equity issues, not just efficiency. If we only focus on efficiency, we create environments that might be efficient but not necessarily pleasant or livable. People want to live in spaces that are efficient, enjoyable, and suitable for their needs.
Jacobsen: Different cultures and subcultures value certain aspects differently, including the acceptability of inefficiency or the aesthetic and feel of a space. That can vary by country, state, or even county. When looking at inefficiencies, do you consider them as positive utilities related to local values? It’s not about inefficiency being inherently negative but understanding the direction and context of those values. Is there a way to fine-tune machine learning to account for these differences?
Odonkor: That’s an intriguing question, and while we haven’t fully implemented this yet, it’s a significant point. Different cultures indeed have varied perspectives on what constitutes inefficiency. For example, in the Western world, power outages are viewed as wholly negative, and we strive to minimize them. However, in developing countries, people may have adapted to power outages as part of daily life. While they may be inconvenient, they aren’t viewed as catastrophic in the same way that they are here.
Instead of telling the machine learning algorithm that minimizing power outages should be the primary objective, we could program it more flexibly. Power outages may be tolerable under certain circumstances, or the algorithm can shift power distribution to account for variables like weaker infrastructure.
You allow the machine learning algorithm to tolerate inefficiencies if they result in a collective benefit. These tweaks seem important, especially considering aspects like energy equity. This flexibility across societies and cities could be advantageous if properly implemented.
We focus on maintaining optimal conditions in most research. However, as you mentioned, there can be utility in so-called inefficiencies, and exploring how machine learning algorithms can leverage these for the greater good is something we are investigating.
Jacobsen: What do people across different cultures typically value when it comes to their vision of a smart city? Clean air, clean water, green spaces—what are the primary and secondary considerations?
Odonkor: That’s an excellent question, and part of the challenge is that there needs to be a universally accepted definition of a smart city. A smart city can mean different things to different people, influenced by what they’ve heard or experienced. In fact, when I teach a class on smart cities, many of my students start out needing to learn what the term truly encompasses.
One recurring theme is the desire for efficiency. People envision smart cities as efficient and livable places. Livability usually means having clean air, reliable transportation, reduced homelessness, and other similar factors. However, one significant concern that comes up repeatedly is privacy.
Balancing privacy and data collection is a complex issue. As I mentioned, my research relies heavily on data captured by various sensors, which is essential for training machine learning algorithms. The more data we have, the better the outcomes. However, generating all that data requires compromising privacy. People want the advantages of a smart city, but not at the cost of their privacy. Finding a middle ground remains an open question we are still trying to solve.
We’ve seen some smart city initiatives struggle or fail because they couldn’t adequately address privacy concerns. A notable example is Sidewalk Labs. I’m not sure if you’re familiar with it—it was a Google-affiliated company attempting to build a smart city project in Canada. It faced significant pushback due to concerns about privacy and data security, and ultimately, it did not move forward because it couldn’t offer sufficient guarantees that people’s privacy would be protected.
Jacobsen: When considering the digitization of everything and the integration of sensors everywhere—visual or otherwise—how do encryption, security, and privacy play into these systems? If the sidewalk has a physical sensor, how do you ensure those are highly encrypted to address privacy concerns? In societies where privacy is not a given right, have some projects infringed on that privacy? This turns into a cybersecurity issue. How do we protect all this digital infrastructure? It may be an open question.
Odonkor: It is indeed an open question. The reality is that we may never reach a point where these systems are 100% secure. We see this even with the most trusted digital systems, such as banking systems, which only guarantee partial protection. However, there have been significant advancements in encrypting data. The solutions will not be limited to smart cities alone; they will likely involve broader applications, such as protecting banking or sensitive health data. These technologies will continue to evolve and be applied in various smart city contexts.
The main challenge is that smart cities are highly complex. They consist of multiple interconnected systems, and any time you have such a distributed system with numerous moving parts, one weak link is enough to create significant issues. Complex systems like this will inherently have weak points, a major challenge for smart cities. However, there is a lot of research in this space. While I don’t specialize in cybersecurity, I am confident that progress will lead us to a point where we are “safe enough” within smart cities.
Jacobsen: Regarding privacy, in some countries, how is this approached?
Odonkor: Yes, that’s an important consideration. In countries like China, for example, technologies that monitor citizens are already in place. While monitoring can have protective and security benefits, it’s also true that these technologies have dual uses. Due to the governmental structure in such countries, it’s easier to install and operate these systems.
In contrast, in the United States, for example, implementing widespread systems like facial recognition often requires more support. So, the global landscape varies depending on societal structures and attitudes toward privacy and governance. We’ll continue to see these differences in how smart city technologies are deployed based on societal norms and legal frameworks.
Jacobsen: What is the most feasible development in smart cities by the end of 2025?
Odonkor: One exciting area is transportation, specifically the electrification of transport. We’re seeing a surge in micro-mobility solutions, such as electric bikes and scooters. On a larger scale, we’re seeing the integration of electric buses and the gradual emergence of autonomous vehicles. An interesting challenge is merging electric autonomous transportation with traditional transportation and navigating the resulting efficiency gains and challenges. This area is evolving quickly, and I’m excited to see how policies adapt to these advancements. Cities are caught off guard by this shift toward electric micro-mobility, so watching how they respond is fascinating.
Jacobsen: Excellent. Philip, thank you so much for your time and insights today.
Odonkor: You’re welcome, Scott.
Jacobsen: Cool, thank you. It was nice meeting you.
Odonkor: Thanks, Scott.
—
Scott Douglas Jacobsen is the Founder of In-Sight Publishing and Editor-in-Chief of In-Sight: Independent Interview-Based Journal (ISSN 2369–6885). He is a Freelance, Independent Journalist with the Canadian Association of Journalists in Good Standing, a Member of PEN Canada, and a Writer for The Good Men Project. Email: Scott.Douglas.Jacobsen@Gmail.Com.
***
If you believe in the work we are doing here at The Good Men Project and want a deeper connection with our community, please join us as a Premium Member today.
Premium Members get to view The Good Men Project with NO ADS. Need more info? A complete list of benefits is here.
—
Photo by Celine Lityo on Unsplash
The post Dr. Philip Odonkor on Smart Cities & Informatics appeared first on The Good Men Project.