Smart Cities: Data Science and the Philippines

Figure 1.1 - LRT-1 near EDSA Station by Bridget Adolfo

Introduction & the Best Part of the Job

Nicholas Paredes [NP]: Prof. Chris [Christopher Monterola], you're a man of many hats. You're a professor in AIM right now and were previously a professor in UP Diliman. You're the school head of A-site. You're a director of Access Lab. You're a National Academician. Is there one [role] in particular that really resonates with you? One that you feel like is a personal dream come true?

Christopher Monterola [CM]: The part of the job that makes me really happy is being an educator. My refuge, really, is science. When I feel that this world doesn't make sense anymore, I go to science and that makes me happy. I would say that the administrative loads are just part of my tour of duty. 

My greatest dream, and my greatest advocacy, is to bring this into the lives of the Filipino professionals. Giving Filipino professionals opportunities has always been the reason for my existence. 

Moving From Physics to Data Science

NP: You started out in Physics, then moved to Data Science. Was that an easy transition? Do you feel like it's a natural progression of things? What's the difference between the two fields?

CM: Well, I’ll say that Physics is a very, very proud field. We feel that if we are in physics, we can do anything and everything. Anything that is exciting is something that we're going to do. I started working on artificial intelligence when I was in college. My dissertation, titled Novel Insights and New Applications of Artificial Neural Networks, was about AI. 

One of my dreams has always been about trying to understand how the human brain works, so I've been doing AI since 1995. That would be more than 30 years now. Even when I was doing my postdoc at Max Planck Institute in Germany, my work was on trying to understand artificial intelligence, the algorithms, and how the brain actually computes information and all. So, yes, it's become very natural for me.

There was no such thing as data science [in the Philippines]. While working in Singapore, working with different industries, organizations, we'd been doing a lot of data science and had been planning to have this [in the Philippines]. 

We felt that this field would come and we were ready for that moment. 

Figure 1.2 - Singrapore's Rail Network Map from https://www.lta.gov.sg/

Smart Cities & Singapore 

NP: Let’s talk about Smart Cities. If you had to explain it personally, how would you define smart cities? Are there key ingredients to the recipe behind a good Smart City? 

CM: I was hired back in 2011 to build a team in Singapore under A*STAR (Agency for Science, Technology and Research). This is Singapore's think tank. A*STAR was made to understand what are the recipes of a smart city, with the general idea being that a smart city through sensors, IoT (Internet of Things), and all the technologies that are available, will be able to make the lives of people within that city easier and better. That is the main idea. That is the goal.

But then there are a lot of complications, in general, when understanding smart cities. Because every component interacts with all other components. They are highly interacting and their interaction is nonlinear. In physics, there are ways to understand when two things collide, what will happen to them. There's the so-called elastic or inelastic collision and you will be able to create an equation out of such. 

But in cities, you're talking about people. People adapt and people form organizations - self-organization. When they collide, the next time around, they will make sure that they will try not to collide and that completely changes the dynamics of your system. So the presence of emergence or self-organization is a key ingredient that arises from these interactions that you need to understand. 

NP: In one of your 2014 papers you talk about how trains are often seen as the backbone of transportation in any urban city. As more and more people start to populate cities and urban rather than rural areas, it becomes imperative to make sure that train and transport in general runs smoothly. Could you talk specifically about your work in Singapore? What sort of problems did you analyze and solve? 

CM: One thing that is good about Singapore is that their ministries have very defined goals. One of the goals given to us included reducing the travel time inside a 15-kilometer distance on average by let’s say, 4 minutes. From 20 minutes to 16 minutes. 

Another one [goal] was to decrease the time-ratio of transportation between private and public transport from 1.6 to 1.4, for instance. That means if you have your own car and if you will be traveling using your private car for one hour, public transport should only be that 1.4 times as long [at maximum]. 

I think in the Philippines, it's about 2.3 or so, which is really bad. We really feel that there is some level of unfairness because the public transport sector is definitely compromised compared to those who own cars.

If you imagine it creating cities, there will be a scaling of energy, a scaling of transportation costs and at the same time, if there are more individuals there, things like pollution and all this inefficiency will also scale up. At the end of the day, this is all about people. Using technology smartly, you can get the feedback of people, and improve their lives.

Transport System Simulations and Human Behavior

CM: In any country, in any location, transportation as a part of life, is among the top three most important issues in those countries. In Singapore, it's second only to foreign immigration. 

In the context of the transport system, are you aware that we are among the Philippines, is actually the first country in [Southeast Asia] to create a train system?

Our MRT was the very first, Singapore’s was the second, but sadly, our transport system is one of the poorest in the region [now]. 

What we did in Singapore was to simulate the movement of all the train systems and how this interacts with the bus system. Simulating this is not an easy feat because you really need to understand how people behave, how people decide every time they go from one place to another.

For example, we learned that people actually do not [always] choose the shortest route. Most literature will tell you that they will be choosing the shortest time, etc., but convenience is very critical to them. In fact, about 30% of the population will not use the shortest route and if you don't include that in your modeling and simulation, you will immediately have a 30% error. So we make sure that we have incorporated that, all the way to understanding how different demographics actually travel from one station to another station.

You need to create some sort of a mechanism-based model where you will very closely be able to simulate the movement of people and for stations when they are compromised. You need to know the reliability of the model for the process of policy.  

Singapore Simulations

CM: I can talk about one of the systems we developed for the Singapore government long before and I think they are still using this.

In this simulation of the entire Singapore area, the first component is typical data science work that allows you to see how long it will take you [to travel]. You can click anywhere on the map that we’ve laid out, start at any location and check how long it will take you to get to another location from that point. 

The heart of this system, however,  is the actual simulation part. This is where we've simulated the movement of all the commuters in Singapore. 

With colored dots, we can show when and how people go to and from the platforms, going to their homes, et cetera. You can click any location and see how many people are tapping in and tapping out [of their commuter cards] and what the waiting time is before they will be able to board from the station. The simulation also shows the bus system.

I can also disrupt stations. For example, if I disrupt Clementi [station], you can see that the disruption in Clementi will propagate immediately to the other transit locations. Even something 15 kilometers away, in Raffles Estate. The government uses this system to let them know exactly where to deploy things like shuttle buses, for example, when there are disruptions. We also added taxis, then private cars later on. 

Predictability of People’s Movements

NP: Is this a simulation that uses live data?  

CM: This is based on actual data, but not live data. It's not necessary to fill it with live data as it happens during that time, because we’ve realized that the weekly patterns of people are quite predictable.

There will be a model drift every now and then, but you can generally forecast what the movement patterns of people are. One of the things that you should realize is that the movement of people is predictable. Everyone will get out of their house, will go to work, and will go back home. Within 90-95% accuracy, if you tell me the three most common locations or the three most common travel patterns of an individual, I will be able to figure out that individual.  

Car Issues in the Philippines and Data Gaps

NP: How much of that simulation model do you think is applicable or adaptable to the Philippines context? Are there cultural differences that you think you'd have to consider or have to anticipate in the Philippines that aren't in Singapore?

CM: There are a lot. The total number of new cars in Singapore since 2005, I think is within the order of 90,000 cars . Their [Singapore’s] goal is to decrease it further because they want public transport to be their choice mode of transport. In contrast, the total number of new cars in the Philippines is within the order of 400,000. [491,000 new cars from 2024 to 2025]. 

You know why that is important in the context of an inclusive and fair city? If everyone commutes, somehow you feel that the world is fair, that this country is inclusive to some point. 

For example, making sure that the people who are working in Makati, have residential locations in Makati must be something that the government ensures will happen. If you are working in one location, make sure that the area where you are working lets you do all the other stuff without commuting outside of that location. The gym should be there, restaurants should be within a walking vicinity, and so on.

Our system is congested and public transport [in the Philippines] is not a choice mode because many [citizens] feel that every time that they commute, the goal is just to survive traveling. 

We had smart data for this [in Singapore]. The data is very granular to the extent that we even have the information of how fast people are walking, moving from one railway to another. We don't have that data in the Philippines. 

Transport isn’t a Utility in the Philippines

CM: Transport, in Singapore, is treated as a utility, like how energy and water are treated in the Philippines. It’s also regulated by the government. In the Philippines, transport is not really regulated and every driver has their own way of doing things.

This makes many things like the scheduling of trains and buses less predictable. It makes modeling in the Philippine setting more complex.

It took us a year to come up with the system [used in Singapore] and this model had ideal conditions, with very good planning & top-notch brains.

So yeah, there are those considerations and you have also to think of the [available] infrastructure. It will be more challenging [in the Philippines] but not impossible.

What Problem/s does Transport Solve and Focusing on Trains

NP: For the Philippines, do you feel like focusing on trains is the correct thing to do? Does focusing on trains generate enough momentum for the other modes of transportation such as buses and jeeps to improve? 

CM: There's a lot of complications when it comes to our transport system. First and foremost, you need to realize that the transport system problem is not about the modes  that will be used, but is simply about moving people from one point to another point. 

If you really want to solve this problem, one of the best solutions, as mentioned earlier, is to make sure that people will be able to access their places of work, relaxation and living and make them as close as possible to their area. Making sure that every area is self-contained is the best solution.

Trains Aren't Always The Best. Taxis aren’t good for traffic

CM: It's also not really about making sure that you will have very good trains. In fact, many studies, [about train transport] say that it is not the most optimal way to move people, because trains will involve creating huge infrastructure that, while you are building it, will cause a lot of traffic. 

And so you ask this question: "What is the economic cost of building and constructing those components?" In many cities, the more optimal way is not really trains but some sort of a tram. It's easier to construct, not as big as trains and is more flexible.

Among the worst in general that cause a lot of traffic in a city are taxis, which are worse than private cars. Private cars allow you to go from one point to another point, but you just stay in that point until the time that you have to go back home. Taxis roam around the cities, sometimes even with no passengers. So, the traffic that the taxis will contribute is actually higher than private cars. So the idea of Grab, the idea that we have this system, that we know where the commuters are, that's very important to understand.

Decentralization & Urban Planning on a Large Scale

NP: Given that the Philippines has a fairly decentralized government, do you feel like it's a smart idea for an urban planner or data scientist to tackle it piecemeal, city by city? Do you feel like a wholly encompassing plan would be a better thing, like a Metro Manila Master Plan?

CM: The short answer is that this should be a whole-of-nation planning. The whole is not only greater, but completely different from the sum of its parts. For example, in Metro Manila, how many cities do we have in Metro Manila? 16.

So you have 16 “Kings” in Metro Manila, and you will not be able to solve the transport problem in Metro Manila by, for example, looking at each of these problems individually. 

Assume that you try to deal with this piecewise; You can, for example, look at each of the cities of Makati, Pasig, and Taguig. By using network science to understand where the traffic will most likely happen in each of these cities individually, you’d be able to find “centers” of traffic within each city. You can predict where the traffic will happen with that methodology, but you will realize that that is not the case when looking at traffic between the three cities together. That, in reality, the traffic in Metro Manila happens at the center. When all three cities are simulated together, the highest traffic center actually completely shifts to areas in between the three cities.

What is my point here? When studied as a whole, the central parts [of traffic concentrations] shift, and the traffic dynamics are more accurately captured. So, you cannot isolate one city from the other city because they all impact each other. People in Makati are coming from all the cities, from south, from north, and if you simply look at Makati alone, it will not work. They are all connected, they all have some level of movement related to each other and they should be [analyzed] together. You cannot look at it piecemeal. 

When you talk about the transport system, it is, again, a connected system, a network, and the existence of the Metro Manila Development Authority is to make sure that all the local government units actually follow the same principles, the same ideas.

The recognition is there, and it doesn't take a rocket scientist to understand that the transportation is connected. Makati has about 300,000 residents, but the total number of people in Makati multiplies by up to a factor of three during work times. 

For other concerns? It's a different animal. If you look at the culture, the way that people look at their life, their world, it might be different. When it comes to the transport system though, the whole is not only greater than, but very different from the sum of its individual parts. So it must be seen as a whole. If you really want to model it.

Shared Infrastructure and Shared Data

Daniel Cotia [DC]: When we're talking about data and urban planning, could we actually advocate for something that is shared between Metro Manila’s 16 cities? For example, having a singular centralized data infrastructure for sensors, traffic technology, for the Internet of Things?

CM: Shared data infrastructure for smart cities among Metro Manila’s cities will definitely help. If you have an idea of the flux of people, the traffic in one location at any point in time, for sure, that will help if they have information about these different locations. It's not just actually good things like the transport system itself, but the negatives like the pollution level.

Things like the carbon emission at any point in time, depending on the wind speed and velocity, has an impact on people's health and that’s critical. There's something called an airshed community or airshed areas. An airshed is different from the usual boundaries [of air pollution]. This is given by DENR. Here, you will know exactly where the pollution is at any point in time.

A good portion of this pollution is anthropologically driven - caused by people, caused by commuters, caused by the transport system. So there is definitely value in looking at this from that perspective and I think that has always been the idea. That's why Metro Manila has a development authority. That is why way, way back, Metro Manila had a governor.

Smart City Initiatives and Context in your Data

NP: Speaking of environmental data, you’ve been involved with projects like Project MINERVA (Monitoring of Indicators for Efficient Redevelopment and Value Assessment) and Project PATURO (Platform for Assessment and Tracking of Urbanization Related Opportunities) in the past, can briefly talk about them?

CM: The only way that you can create and make a difference in a smart city initiative is if you have a government, an LGU, who can actually impose and implement and sustain the efforts.

Project MINERVA is in Baguio [City] and Project PATURO is in Cauayan, Isabela. If you’re going to look at the urban works that we did in the Philippines, I think it's good to highlight Baguio City [and Project MINERVA]. Maybe one of the things I would like to emphasize there is that you don't always need a sophisticated model. Sometimes all you need is data. 

For example, one of the most impactful things that happened in Baguio was when we were looking at the water quality of the different barangays based on our sensors. We couldn't explain why it is that when it rains, there were certain barangays where the pollution increased. The content of their water was different. Then, Baguio Mayor Magalong told us that the reason for that is when it rains, citizens open their poso negro to clean it when it rains heavily. This was something that compromised the entire population. 

There was no complex modeling that was done for that. It's just that all of a sudden, when you take it into context and talk to the locals, you gain insight. 

Open Street Map & Policy Making

DC: Open Street Map is used and highly advocated for in map making and urban planning. How can Open Street Map and other open source data sets actually help out planners, especially when we're talking about simulations and data science?

CM: We’ve used OpenStreetMap extensively for many purposes. 

We utilized it to forecast the movement of people. Who are the people who will be riding in this train station or in the bus station? And the forecasting was quite accurate. 

There is this idea of familiar strangers, which is something I learned in Singapore, but I think is taught globally. There is an almost 80% chance that if we are on the same bus today, that in a week, we'll meet again. So, we're actually there in the same space, we don't know each other, we don't talk, but we're always meeting and we have that transient experience. 

By looking at the Open Street Map in the Philippine setting, with the help of other things, we were able to make a very accurate forecast of the movement of wealth in the Philippines. It helps us understand which municipalities will eventually become second class or first class municipalities in years to come. We’ve also helped the government determine where we should be putting schools so that this school will be accessible to the different students who are in compromised areas. An interesting finding that we got from our work is that one of the indicators of the movement of people is coffee shops! 

So, if our urban planners are serious, if our architects are serious, and if our policy makers are serious on really creating something innovative in the Philippine setting, the best idea would be to expose themselves to all these technologies.

About Prof. Chris, PhD:

Dr. Christopher P. Monterola, is a distinguished physicist, machine learning scientist, and educator with a prolific academic and professional track record. As the founding head of the Asian Institute of Management's Aboitiz School of Innovation, Technology, and Entrepreneurship (ASITE), Dr. Monterola is at the forefront of advancing interdisciplinary education and innovation. He also leads the Analytics, Computing, and Complex Systems (ACCeSs) Lab as its Principal Scientist and Executive Managing Director.

from ChrisMonterola.com