The Manhattan Project

Budhendra Bhaduri's Interview

Printer-friendly version

Budhendra Bhaduri

Budhendra “Budhu” Bhaduri is a Corporate Research Fellow and group leader of the Geographic Information Science and Technology Group in the Computing and Computational Sciences Directorate at Oak Ridge National Laboratory (ORNL). He has worked at ORNL since 1998. In this interview, Dr. Bhaduri describes how his group researches global population dynamics, including studying population distribution and movement from rural areas to cities. He explains how Oak Ridge became involved in population research, and expresses his hopes for how geographic data can be used to improve living conditions for people around the world.
Manhattan Project Location(s): 
Date of Interview: 
April 25, 2018
Location of the Interview: 
Oak Ridge
Transcript: 

Cindy Kelly: I’m Cindy Kelly. It is Wednesday, April 25, 2018. I have with me a scientist who is working in the Computing and Computational Sciences Directorate. My first question to him is to say his name and spell it.

Budhu Bhaduri: My name is Budhu Bhaduri, spelled B-u-d-h-u, and last name spelled B-h-a-d-u-r-i.

Kelly:  Thank you. My first question is to tell us something about yourself. Where were you born, what was your childhood, how did you become interested in becoming a scientist?

Bhaduri: I was born and raised in a small city in the eastern part of India called Calcutta. I always get that reaction because people feel that it’s really, really large. It’s a lot of people, but geographically speaking, it’s very, very small, and that’s my area of research. I always like to point out that distance dimensions are very relative, how we think what is small versus what is big. There are a lot of people and lot of concrete and asphalt, but the space is quite restricted. People are packed in.

I came to the United States in 1992 to pursue graduate school, particularly motivated to do something pursuing environmental and applied research. I was always inclined to science. I liked science when I was going to school. I remember distinctly they used to have a one-hour program on National Geographic every Sunday afternoon. I was sort of hooked to that one, and there was this particular one that talked about cleaning up streams in Germany that somehow stuck to me. It was one of those first big push in the sail in that direction, saying, “You know, I would like to be part of something like that, making a difference for the planet.”

That’s how I got into science. I studied geology in college. As I always tell people, I started at the center of the earth, and then started floating upwards. As I went through school, I started floating up and up and up, and then in graduate school, I started doing urban hydrology, so I was looking at streams. Then for my doctoral work I started looking at satellite images and how those satellite images start telling us things about the planet. I went all the way from the core to space, and I think that’s where I had to stop and start working.

I came to Oak Ridge after graduate school back in 1998 with a plan of working here for two years and then going somewhere else. I will have to admit that it’s still that plan, a two-year plan. I’m just terrible at executing it. I guess when you’re having this much fun, you just don’t have enough time to execute your two-year exit from Oak Ridge. One of these days, one of these days.

Kelly:  That’s great. Well, tell us, what’s been so much fun? What have you been working on?

Bhaduri: I came and joined this group that is called the Geographic Information Science and Technology. Today it’s mainstream societal technology that everybody carries in their phone, particularly the GPS [Global Positioning System] and navigation and maps. If you look back, you know, the group had the word “geographic data” in its name as early as 1972, which is pretty pioneering. I had a curiosity about learning, why does a national laboratory that has been set to pursue nuclear science and engineering do geography?

What we found out is that during the Manhattan Project, there needed to be a way to disguise the location of where these three plants were. When you look at the old maps of the reservation, you will always find that they had a tilted north arrow, which is unacceptable in geographic or cartography research. The north arrow is always straight, pointing up. The reason is that all the three sites, Y-12, K-25 and X-10 at that time, they had their own projection or coordinate systems. If you did not know the magic conversion, even if you had the map, you could not actually get the latitude and longitude of these places. That’s how people were brought in to create what I call geographic data in disguise. That was one of the defense or security mechanisms.

One of the work that the group has done very successfully over the last three decades – starting last three decades – is we are uniquely recognized across the world for our work in understanding global population distribution and dynamics. Where do people live and where, how people move on the landscape for a variety of reasons, after disasters, for geopolitical unrest. Even things like transportation on a daily scale, how people move within a city, because the movement of people actually drives our need for energy. If we didn’t move or if everybody stayed home, we would not have the societal challenges like congestion and pollution. Life would be very simple.

You know, historically, the reason Oak Ridge became a focal point of population research is because of a requirement from the Nuclear Regulatory Commission that stipulates all facilities with nuclear material to understand in the decadal census years how many people live within fifty miles of these facilities. Across the nation. Anywhere you have a nuclear reactor, for example, in universities, they have to do this as well. When we did not have geographic information systems or software that allows us to do this almost on our phone today, it was a big scientific experiment. People had to get a lot of data together and come up with a map that showed where people lived or how many people might be. This was done usually through census data. From that type of research, eventually it turned out that we could actually understand a lot about how this planet is populated.

It’s a very interesting problem, and it provides sort of the backdrop of why satellites became so important. If you think about the world’s seven and a half billion people and you make families of four, which is about a standard assumption of a household, two parents and two kids, and you give each family about a quarter of an acre, which is quite a bit of space in the global standard, you can fit every family in Alaska and still have 10,000 square miles left. If you give each family about an eighth of an acre, which is a little over 5,000 square feet, which is pretty luxurious space in the global standard, you can still fit everybody in Texas and Oklahoma without stacking people up.

The challenge is that there is enough space on this planet and people are not concentrated, so they are distributed. When you are trying to understand where people are, it’s really looking for a needle in a haystack, because only two to three percent of the land surface is actually populated by humans. The amazing thing that we found out is that when you have satellites and they are taking pictures every, you know, all the time at unprecedented frequency, you can see a lot. You can see a lot much better. Now we have pictures that are a meter or half a meter resolution, so you can see very detailed features on this landscape. Finding those is a very important scientific challenge.

That’s where the computers came in. It’s a task that humans cannot perform by just using eyeballs. If you think about a one-meter resolution picture of the 48 states, that’s about 10 trillion pixels. If you had humans go through 10 trillion pixels looking for where there is a structure, it is almost impossible. That’s where computing comes in. You want to take advantage of the big, large computers that are very good at sifting through data, and use scientific techniques like machine learning, artificial intelligence, deep learning, to let the machines find this human-built structure. Nature never builds a building. If you see a building anywhere, you know that some humans must have been there. We are using those structures as proxy for the human presence.

The amazing thing that has happened is, as we are trying to look across the planet, especially in the Global South, south of the Equator – places like Africa, Sub-Saharan Africa, in the southeast regions of Afghanistan, Pakistan, Kyrgyzstan, the entire region – we do not really know where people are there, because those official censuses have not happened. Let’s take Afghanistan as an example. The last official census in Afghanistan was 1979. It’s been a long time since then and we don’t really have a good understanding of where people are and how many people might be in Afghanistan. Nigeria is the same example, where the last official census was in 2006. The last official estimate was 2012.

About four years back, I had an opportunity to intersect a group from the Bill and Melinda Gates Foundation. They are interested in eradicating polio, or that’s their mission, to eradicate polio. The task is fairly simple at one level, which is find every child who is between zero and five years old and give them the vaccine. But when you have your last official estimate done in 2012, the people you are trying to find do not exist in a database. The only way you can find them is when you can find each and every settlement or individual structures and huts in Nigeria. We started working with them and we are building a bottom-up estimate of Nigerian population, and a lot of countries, for that matter. Afghanistan is one of them.

I call it this amazing achievement of finding the missing millions, right. These people have never existed for our civilization, which is remarkable. As human beings living on this planet, we have a moral, ethical, almost responsibility to make sure that we find them, and we ensure their wellbeing, right, so the human security and the quality of life.

It all goes back to what you first started saying, these sort of Manhattan Project-scale investments as they have come through decades, how they are making an impact today for society. We would not be here unless we had to first look for people within 50 miles of every nuclear operation to ensure their safety.

We are even doing interesting research – for example, if you wanted to create nuclear power plants for the future, you have to ensure that you find places on the planet or even in the nation where you can safely create a nuclear plant. Which means you have to know where people will be in 2030 or 2050. One of the science projects that we have done is to understand that if things worked just like they have worked the last 30 years in the 48 states, where will people be and how many people will be there in year 2030 and 2050? You can plan ahead of understanding where, you know, how feasible it is. This has impact in many of our nation’s security. If you know the airports, where they are, they’re all getting surrounded by residential developments, and that has a huge impact in terms of noise pollution and abatement, right. Knowing where people are.

One of the key research that we do in population dynamics at Oak Ridge is understanding where people are when they are not sleeping at home, which the Census reports to us. If you look at that, that’s only one-third of the 24-hour cycle, right. So maybe eight hours we are sleeping at home. The rest of the time, we are somewhere else. Knowing and understanding and documenting, creating datasets, tell us where people actually are during those other sixteen hours of the day. It allows us to better understand what kind of environmental conditions they’re exposed to. What type of services do they need to make their lives better? Do they have access to jobs and healthcare? It all goes back down to where people are. If we don’t have to worry about people, we won’t be doing any research.

At the end, our whole research enterprise is trying to make humanity better, trying to make the quality of life better. And not just for the United States, but we want to have a global impact. That’s the fun part that has kept me here for twenty years.

Kelly:  Oh, my goodness. Well, it’s incredibly challenging.

Bhaduri: It is.

Kelly: It must be.

Bhaduri: It is.

Kelly:  Do you work with sociologists and anthropologists and historians? I mean, the migration of people is very complex.

Bhaduri: I work with a very interdisciplinary group of people. Scientists who have expertise all the way from social science, for example, to geography, anthropology. All the way to people who work with very large computers, a lot of data. Like electrical engineers, computer scientists, computer engineers, civil engineers, transportation engineers. Because unless you bring a very interdisciplinary approach, some of these societal challenges are impossible to solve. If you just had a machine who could look at images, you will not understand why people do what people do. Explaining why people are where is an important part of it. Yes, we have to constantly work with an interdisciplinary group of experts.

Kelly:  I think it’s Thomas Friedman who talks about the greatest migration that’s happened in the last ten years from – just take China and its migration from rural areas to the cities. Are these the kinds of things you’ve been —

Bhaduri: That’s a very important trend that we are very aware of. Cities act like a magnet. For the last four years I have been directing an institute at Oak Ridge National Laboratory called the Urban Dynamics Institute. We are looking at this global phenomenon of population and land use, water and energy, sustainable mobility and then urban resiliency from the perspective of data and computing what sort of insights can we bring out. One of these trends that we can now very well observe and monitor across the planet is this urban migration.

Eventually, you want to know how the planet is doing. Today, you can wake up in the morning and take out your phone and immediately see what’s outside or how hot it is outside or is it raining outside, so you know what’s happening in your immediate vicinity. But wouldn’t it be nice to know what is the pulse of the planet today? If I split the planet in one-kilometer cells, I would like to know the health of the planet today. What’s happening in every one of those one-kilometer, one square kilometer areas? Are there shortages of water, are there excesses of water, is there flooding going on, is there a drought going on?  

That type of information flow we can clearly think about getting towards, because we now have this incredible amount of dataflow from satellites, from sensors that are out there. Then we also have access to these large computers that can allow us to take all these data streams together and help us create models that generate information that we use.

My best or most favorite example is the weather app. There are two apps on our phones that have truly attained in my view a globally pervasive status. One is navigation. So we all have an application that tells us if I wanted to go from one place to the other, how do I go there, how the traffic might be.

The second one is weather. If you think about that weather app, it gives you a number and an icon that either shows the moon or the sun and a little bit of cloud and the rain. But if you really think through hard saying, “How did I get that number?” It starts with a lot of satellites collecting a lot of data that gets integrated by scientists and processed through very large computers, and then at the end, all you see on your phone is a number and a picture. We are making life-changing decisions or life-dictating decisions based upon that. We often forget that, how robust and pervasive science that is standing behind to get us that service.

There should be many, many more of these kinds. As I talk to my colleagues, I find out that they would like us to know what the water quality is. If you drank water today, did you check your phone to see how the water quality is in the tap of Oak Ridge, Tennessee? I’m sure a lot of people worry about it. We all worry about water or the air. Information makes us live a better life. I think the interesting, but challenging, part is how do you make all those things come together and create these benefits for the better, bigger social good?

Kelly:  Well, you’re making me think about a lot of things. What about the trend that has been predicted, that the seas will rise and the coastal areas, which have been the most attractive for people to live in, may become flooded? Has your group or others taken your data to kind of analyze what’s going to happen in the next —

Bhaduri: We know that there is a clear trend that in some parts of the world the sea level is clearly rising. There are coastal areas that are getting inundated, so there is a displacement of people. We know for a fact that when you have a devastating flood, it brings displacement of people every year. Think about the 2017 hurricane season, you know, what happened in Houston. You will always have displacement of people from flood.

I think where we are not quite mature is to understand exactly where those people will end up going. Are they going to just move a little farther inland, or they will take a longer migration to go back to other places? I think the social scientists are clearly interested. My social science colleagues are clearly interested in that phenomena. Collecting those kinds of datasets are not an easy way of doing it.

A lot of the data that gets reported is because of the popular press. Anytime you have a hurricane and people are moving, you find out how many people lost their homes and how many people had to be relocated in Colorado. But as time goes and the newsworthiness of that event sort of dissipates, you don’t really quite hear the news saying so many people moved back to New Orleans from Colorado, because that’s not quite as interesting news to report. These kinds of datasets quickly get out of the public space, and we don’t have a very good way of going and grabbing those things to really understand the nature of the migrations.

There’s quite a few researchers who are trying to understand and collect data, but not at a global scale that we would like. Observation-based understanding from data is one way of getting towards that answer. If you see refugee camps popping up somewhere or informal settlements popping up in the fringes of the cities, those people must be coming from somewhere. That’s how usually the developments work. Your people come in. It’s a matter of equating, can you connect the dots and actually understand which areas are people migrating out of and which areas they are likely to go.

It’s a very complex social phenomena. There is not just a very simple thing that you can look at and say—there are lots of informal settlements in very large cities. If you go to Southeast Asia, you can see a lot of them, and they absorb a lot of these refugees without building new structures. So you will not always see a one-to-one correspondence of people being  displaced and homes being built. Because they will just go in and increase the population density of those areas. But it’s one way of looking at it, what you can see happening on the surface of the planet.

Kelly:  Do governments sort of try to reach out to you and learn what you have learned about their cities or their migration patterns in their countries?

Bhaduri: I think our job is to help. We work with the educational research community and also the federal government, the U.S. federal government, and agencies that work internationally. I think where we are making the biggest difference is telling them where people are and what sort of conditions these people are living in. Many developing nations have initiatives to improve the quality of life of people they feel are living in slums, which is a very crude way of describing people living in informal settlements without having access to a lot of the societal services. Because what we might call slums in the United States will look like palaces in some other country. It’s a relative description.

But a lot of these countries don’t even know how much slum areas do they have and how many people live in the slums. If you do not know – in our science we call it the “basement problem.” That means you do not even know the denominator. You may have a $100 million to improve the quality of life of people, but if you do not know how many people, you cannot start any plans.

It’s a fundamental foundational data that you need to first get before you can make the first ever progress or take the first step. We are helping organizations understand where these sort of people are, and you know, do you take services back out to them, or what would be the best way to serve these people with the societal services that they may be entitled to from their governments.

Kelly: Well, it’s immensely complex and very important work that you’re doing. Thinking about the immigration problem from the Middle East to Europe and how factors have changed radically in the last five years with the nation-states being welcoming and then not so welcoming. And now putting pressure on the states in which these people are leaving from to block even their passage across the ocean.

Bhaduri: It is complex. Humans as a species have clearly demonstrated an affinity, a trend and a commitment, to move from the day we came on this planet. The question is why and there could be many explanations of why humans try to move. We have moved across the planet and we have gone from East Africa, now we populate most of the land mass in this planet. Some of which is natural and some of it is not. Some of which has come from policies that nations have taken.

We have created countries, we have created divides, nations have created divides by their own. The number of countries keeps going up as we divide, and that creates internal migration. If you look at the Indian subcontinent, those three countries are an outcome of a political settlement that the British had with the Indian government. If you look at Sudan and South Sudan, that’s an internal agreement that they agreed to create two countries out of one. I don’t see this stopping. If you believe in the human history, these kinds of phenomena are only going to be observed over and over and over.

I think what’s more interesting is that the mobility has gone up. Accessibility to place is clearly easier. There is a recent Nature paper by some researchers at Oxford and Google that shows that about 80-plus percent of the human population are within two hours of a city of 50,000 people. Accessibility has gone [up] quite a bit.

The revolution that we are seeing in urban cyber infrastructure is also very interesting. If you think about our research, we have always created this broad divide between who’s urban and who’s rural. If you asked people what’s the underlying assumption, it is essentially population density. You take a small city or a small place like Cleveland, Tennessee, which is a city by a legal definition, because they’re incorporated. It’s about 40,000 people, but they have faster Internet than anybody on this planet. They can stream movies [faster] than you and I, they can order commodities from the Internet shopping platforms and get it delivered at the same time. Are they any less urban than people who live in a big city?

These sorts of definitions are changing, so the quality of life is changing in other different ways. Which, again, brings it to the point of what is the rule of computing as a field – which includes cyber and cyber security – in impacting human quality of life? We are almost, in some ways, changing what definition of who is and what is urban versus what is rural. Those are sort of, we are blurring that, which science should do. Science should change the way we think about our surroundings. That’s the ultimate success of science if we can make this positive impact on humankind.

Kelly:  Is your team well-supported? Is it growing, is it —

Bhaduri: Our team is growing. This is sort of a new area of research that has emerged. You know, fifteen, twenty years back, if you asked people to think if they could get directions on their phones, or popping up their browser and look at any place on earth, what it looks like through a mapping interface like Google Earth or Bing Maps from Microsoft, that would have sounded like science fiction.

I always like to remind people the story about how geographic data has given birth to computing and maybe it’s relevant in this context. The first census in the U.S. was back in 1890 [misspoke: 1790]. The mechanism was people, marshals on horseback, went around the country and counted people. They would come back, tally the numbers. Increasingly, if you look at the trend in census, the kinds of data we started collecting went beyond the basic, how many people and a gender and age. We wanted to know which religion they followed, what kind of socioeconomic strength they had. Even terms like “pauperism,” which we don’t use anymore. We call it poverty today. But pauperism is a word that is documented in the census that people wanted to know.

The real challenge came in a later part of the twentieth century – in the 1940s. The estimate was that there was to be 65 million people counted in the next census. What they realized is that if you repeated that process of horseback riders going and counting people, it would take you over ten years to complete that process. Necessity is the mother of invention; people started looking for other types of data or solutions. They found this biostatistician in Baltimore, Maryland, who had this technology of very fast counting. That was the punch-card counter that was successfully used to count 63½ million people. The machine was blamed for the missing one and a half million. Between 1911 and 1924, I believe – I have to check – but that company was successfully commercialized, and between two mergers, it created IBM. Geographic data drove the birth of computing, because you had too much data to crunch.

We are seeing another version of it today with so many satellites out in space collecting data. Sorry, the first census was 1790, so it was 1890, a hundred years later, when we had that big challenge. Then 1911 to 1924 when we took the leap of faith with technology and computing. It took about 100 years to hit that problem space of too many people to count with people on horseback. We are seeing another sort of data revolution, and that would give birth to new generations of computing.

It’s not hard to understand. We all live this life, and I have the fortunate opportunity to have fun with everyday life, which is pretty unique.

Kelly: That’s wonderful. Would you advise other young people to go into science?

Bhaduri: Oh, absolutely. You know, especially — it’s an interesting question — the first answer is absolutely. I would like to inspire everybody to be part of science. Whether it’s physical sciences or social sciences, people should aspire to become scientists. The interesting thing I will point out, which in my life experience and over my career I have struggled with, is I have successfully lived a life of a scientist and a researcher within a group that is pursuing research in geographic information science. Although we as a nation have lumped geography largely into the school of liberal arts. When having three kids going through high school – one is just out and two are still in high school – when I think about their life or what sort of counseling and guidance they get, I don’t think they are being told, “You are so good in science. You should go study geography.” Right?

I think the time has come when these sorts of new scientific areas are emerging and exciting everybody, right. How many people do you think are going to give up their navigation app or their weather app or the GPS in their phone, right? It is a very strong aspect of science and the next generation. That is also our responsibility as scientists and researchers, to inspire and motivate and then grow the next generation of scientists and engineers as part of our workforce. Absolutely, it has to happen.