In the spring of 2017, I was living in Bhutan, a small remote country enclaved between two giants, China and India. I remember traveling from the closest airport, in Jakar, to the village I was living in, Lignmethan. It is an entire day driving up and down, valleys and mountains passing by, through the “highway”—a mostly unpaved two-lane road with no markings, largely no safety rails on the sides, and boulders and road bumps requiring excellent driving skills to navigate. It is such an incredibly beautiful landscape, and literally one of the most remote places on Earth, due to limited visa permits and long travel times.
I was working for a hazelnut company (‘Mountain Hazelnuts’), a for-profit fully foreign-invested social enterprise that provides small trees at no cost to thousands of low-income farmers, to more than double their income reliably for years selling back to the company what they harvest, while the roots of the hazelnut trees protect the land from erosion. The company is based on the coveted “triple bottom line” ([[Notes#^6]]) model of sustainability: for social cause, for profit, and for the environment. It is also a company that loves data and measures everything. I worked with the company for a few months helping to figure out their logistics for the collection of initial harvests from millions of young hazelnut trees, scattered across the country along those infinite roads. I remember being struck by the gorgeous view of the natural landscape, by moments of awe for the beauty and privilege to be there. At the same time, I wondered to myself if it made any sense for an astrophysicist like myself to be doing this type of work.
I had studied solar physics as well as completed a doctorate degree to explain magnetoacoustic plasma waves. Following that, I had worked with NASA rockets and satellite missions. My life had not involved many hazelnuts, besides eating them with complete disregard for the paradises they might have come from. Neither was I an expert in logistics, or even in Bhutan. I would have needed some time to locate the country on a map just a few months before.
On the sunny day of my arrival in Bhutan, riding in the car from the airport, I could talk for hours to the driver about the origin of the light that was shining over the landscape and the oddities of the plasma waves up there, still so mysterious to the few people that dedicate their lives to understand them. And to be honest, the driver did love astronomy, and we had plenty of time to burn. It was a beautiful, chatty long ride, but I wasn’t there to talk about astrophysics.
I had gone to Bhutan to test the entire concept of this book, the idea that the value of a scientist lies more in their acquired skills and experiences than the facts in their heads. In my years of training as a “professional understander” of physics, I had gained a few potent tools: these included advanced math, modeling, the methods for creating and testing falsifiable hypothesis, and computer coding. These tools proved useful for understanding the distant sun, but—as my hypothesis went—could equally work to offer substantive help closer to home on Earth, for example nut logistics in Bhutan. In fact, such specialized tools I had learned could—and should—work on pretty much any topic measurable with data. The question is, then, how to redirect their usefulness towards maximum positive impact in society versus limiting ourselves to mostly the measurement and understanding of processes.
The company I was working for is extraordinary. Indeed, it is profiled in a Stanford Business School case study due to their broad use of data as part of a strategy for sustainability. I knew I had more than enough material to work with, along with many questions they wanted me to help answer. Having sufficient data to inform strategic decisions is pretty much the rationale behind data science. So I was also wondering if me being in this remote area of the world was just another case of a data scientist working from afar, but within the same framework. My official role was to help create a model to estimate harvest volume and devise the logistical operations to collect it. Yet, I believe there was a significant step beyond the data processing part that made a lot of difference. As we stopped for lunch, I was graphically reminded that implementing scientific solutions in society through data should go beyond just numbers, and it is not just a paternalistic one-way street.
As you approach the door of the restaurant in Sengor (within the Mongar province, atop the last big mountain pass of our trip), you are greeted by the view of two five-foot-long drawings of erect phalluses on the side of the entrance. As you come in, another wooden phallus figure hangs high and erect above on the wall, next to a picture of the king of Bhutan and a calendar. Turns out that in Bhutan, especially in the remote eastern provinces, this is an old Buddhist symbol to protect houses from demons.
Momentarily providing a bit of culture shock, these symbols helped me to start unfolding a deeper understanding of the Bhutanese way of life. The phallus was a conversation starter. I learned about Lama Kunley, who introduced Buddhism to the country in 1499. Other Buddhist practices followed in Bhutan include always going clockwise around stupas (religious stone figures, sometimes in the middle of the road) or the absolute veneration for the sentient life of all animals and insects. The religion forms a strong link to their culture, with a strong respect for nature, low levels of corruption, and an increased feeling of belonging together. The country is, moreover, undergoing a series of profound changes of modernization, from turning an absolute monarchy into a parliamentarian monarchy in 2008, universal healthcare and primary education, or the very recent technology push where TV and Facebook almost arrived at some places at the same time. I also learned that despite great gaps in adult literacy, virtually everyone uses voice messages and pictures with the apps on their phones to communicate, such as WhatsApp or WeChat. The Bhutanese people create a closely linked community with virtually two degrees of separation to anyone else in the country([[Notes#^7]]). All of these factors play a crucial role defining reality for the Bhutanese people.
These cultural points might seem outside the scope of my specific work in measuring and assessing hazelnut harvests, at least at first glance. But trying to impose a data-driven solution without such cultural context can definitely backfire. That is the case, just to name an example, of an infamous trend in lake Malawi since 2015. Very well-intentioned global health workers have provided millions of free insecticide-treated nets to prevent malaria from spreading. These are mostly to be dragged to a nearby lake and used for fishing. Not only did that community continue to experience high rates of malaria, people continued to lack sufficient access to malaria prevention, and now they also had toxins in their lake and a dying fish population[iii]. Beyond data, perhaps less easily quantified factors such as religion, cultural and social norms, and access to resources need to be considered. Given the growth of data and information available as technology evolves, scientists—who are trained to understand data—are better positioned to help find certain parts of those solutions that will have an impact on reality.
I came to understand very quickly that the remoteness of Bhutan, the character of its people, and the lack of data scientists available to the region would yield my solutions stalled as soon as I left—or worse, void should any tweaks be needed. I was, however, fortunate to have knowledge of previous cases that had failed due to lack of cultural understanding. I was prepared to learn more about the history and religion of the place in which I had just landed. I quickly grasped a need to become familiar with the quirks of driving directions to account for the literacy constraints of some of the farmers and the remoteness of financial institutions, or the tree orchards themselves from roads that trucks could navigate to.
The particular solution we found is not the important part of this book, rather the approach regarding how the scientific skills were used. Over the time I spent there, we quickly identified both the strengths and the weaknesses in our data. We used the regions with the best data to start testing and built a plan to begin completing the data in the rest of the system. We also created a quick prototype of the model to test for blind spots. We found some, such as dealing with non-connectivity zones or being very mindful of the cost of data connectivity. We created a data science program to transfers the skills needed to understand the solution and the tools used. To ease, and constantly test, the knowledge transfer, we established a continued supervised learning among students; I would only teach different parts of the curricula to each student, and then they teach each other the rest while I listen to their lessons. We did some experiments to test travel times, and logistic recommendations. We were at most two phone calls away from speaking directly with someone who knew first hand any direction or location across the country. The Bhutanese are a closely-knit community. By the time I left, we had a first set of logistics recommendations and I left with a fair confidence that the work could continue to improve once I had left.
Focusing on impact meant observing a more complex dynamic as the data tell on paper. It meant making things more complicated; it meant stretching the tools of science to include these constraints; it meant selecting what to train on, how to train, what to build, what tools to develop; and what hypothesis to test and build on. And then iterate and adjust as soon as possible, driven by pragmatism of the messy reality, not dogmatism on top of partial data. It might mean going against what the data says because other factors might be more important, politically or religiously. Impact science means leveraging a scientist first for the skills they have, not the knowledge. Skill-based value, versus knowledge-based value. The knowledge is useful, but more useful, and universal, are the skills learned to create that knowledge a scientist might have.
I remember hearing an interview where the host asked a famous scientist what is the one bit of information, one and only one, you would choose to represent how advanced humankind is. If you had one message to pass on just before ours disappears as a legacy for any future civilization. Or if you had one message to place on an interplanetary spacecraft that will probably survive us before it reaches an alien civilization. Some scientists, if asked, might choose a basic factoid of their field, like the atomic nature of reality, some equations of physics, the axioms of mathematics, or a definition of life. I struggle with this question. It occurs to me if our civilization disappeared, we would forever lose our beloved novels, poems, songs, legends, stories, experiences, religions, or moral values. But not science. Whichever civilization emerged from the ashes, they would recover every single physical law, fact, and theorem of science. The same ones we have to the last decimal point. Sooner or later, but exactly the same. Maybe other names and units, but we would recover the knowledge and concepts of nuclear physics, relativity, bacteria, vaccines, or fluid dynamics. That’s the beauty of how universal science is. That body of knowledge would re-emerge identical, like a phoenix, from the right set of skills, like the scientific method. More interestingly, whoever can create the right skills first would create the knowledge that might define a civilization, and in turn further explore the fruits of those skills. Science profoundly influences reality: theory of germs, atomic bombs, the steam engine. Moreover, reality then could be different depending on who discovers what first, and for what purpose. Then it wouldn’t be just the facts that defines science’s contribution to the world—it would be society that shapes and defines what direction science goes, which in turn shapes society. When I think of that quiz, I wonder how our scientific knowledge, and our world, could have been different, should our history had been slightly different. What more would we know of, because it was important or relevant to the ruling power, and what would we not know? Would we know more about astronomy if the Mayan empire had not collapsed? How would science be if it were not for the European Middle Ages that halted the liberal renaissance revolution of science? How many Einsteins have died victims of hunger, poverty, or bigotry, and their discoveries unknown, because they were not born in the “right” place, at the right time, or with the right skin color, gender, or gender identity?
It is absurd to think science and society do not define, and depend on, each other’s potential, even when the facts of science are universal. It is absurd to think science is just about data, discovering facts, and collecting an ever-growing mountain of knowledge wealth. The focus of science should be on its ability to create impact in the society it lives in, that it helps define, and from which it depends, financially and morally. For that we do need the skills of science, and we do need a growing harvest of knowledge, but we also need many other parts of the hard puzzle that makes real impact. Is not only knowledge, it is the whole puzzle. When we succeed, we save the lives of millions, we explore other worlds, we transcend our limits, create new skills, and discover new dimensions of knowledge. When we fail, we roll back the clock of progress, we threaten the environment we depend on, we risk eradicating our own existence.
Moreover, we have historically never seen such cadence of progress, such revolution of technologies, skills and potential, and responsibility of science and technology, as we do today. The chairman of the World Economic Forum, Professor Klaus Schwab, coined the term “[iv]fourth industrial revolution” to refer to it. The first industrial revolution was the mechanization using steam engines. The second, mass production using electricity. The third, the digital revolution. The fourth is different; it is not the effect of a revolutionary technology, but the compounding effect of many at the same time—blurring the lines of the physical and digital worlds, machines and humans, or our social contract of work, governance, or identity. In this context, data, data science and science in general play a fundamental role. Not by themselves, not by the knowledge and skills they create, but by the interrelation with other factors. It will be our capacity to understand the impact of data, data science, and science in general, that will help us shape this new world.
To help explore this relation between science and impact, we can use some hard lessons when the gap between understanding the data and creating a positive impact was too large to bridge. I have chosen a few recent cases of such failures. I have also chosen a few recent cases of successes. My hope is that going forward with these, exploring their context and drivers, and raising awareness of this framework can motivate and help create more impact science and more impact scientists.
Thank you for the interest reading this far!
This book is available for purchase on Amazon. You could also go there and give it a nice review ;)
For feedback: [email protected]