One Better

This is an article from the fall 2016 issue of LSA Magazine. Read more stories from the magazine.

In much of the developing world, if you want to know which villages don’t have electricity, you hire a guy with a motorbike and send him into the countryside. If he finds villages that don’t have access to the power grid, he’ll tell you where they are.

It’s not exactly efficient, but it is straightforward. What’s harder to understand is when a village definitely is connected to the power grid, and the electricity still doesn’t work.

These are two of the problems Brian Min, an associate professor of political science, tackles in his work monitoring global energy access. “Electricity is the lifeblood of the modern economy, which is why energy access is so important,” he explains. “Around the world, 1.2 billion people have no electricity. It’s one of the biggest obstacles to alleviating poverty.”

International aid organizations have made huge investments in programs designed to improve energy access for these communities, but evaluating how effective these programs are—and whether or not the money is going where it is supposed to be going—can be difficult.

“These groups spend money to extend power lines to some rural village, but does the electricity keep working after project teams have left?” asks Min. “Is anyone benefiting? No one knows.”

To begin to answer these questions, Min has taken a step back—way back—to look at the Earth from the perspective of satellites operated by the National Oceanic and Atmospheric Administration (NOAA). When electricity in a village is working, light is visible in the darkness of night, while places without power appear dark. Satellite data provide, for the very first time, a kind of map that illuminates where electricity is actually working.

“With this information,” says Min, “we hope to create a much more focused and rational approach to improving energy access.”

In the Dark

For years, NOAA has produced an annual composite of night light images around the world. The images are clear and crisp because they’ve been filtered to remove clouds and other obscuring atmospheric disturbances, a process that also, unfortunately, strips away the dim light produced by irregular electricity in much of the developing world.

So Min asked NOAA for access to its entire archive of raw orbital imagery over India since 1993: five terabytes of satellite images containing tens of thousands of orbital image strips collected over some 8,000 nights. Working with U-M’s Advanced Research Computing consultants and using Flux—U-M’s shared, high-performance computing cluster—Min churned through these images and extracted the light output for more than 600,000 settlements in India to construct a database with nearly 5 billion observations.

It was a huge computational effort, Min says, but the harder part was distinguishing electric light from light coming from other sources, such as reflections from moonlight or electronic static from the sensor itself. To address these issues, he used background noise reduction, a technique he borrowed from astronomy.

To identify an actual light source, astronomers must first filter out noise. Looking at what they know is a dark, empty part of the sky, they can measure the level of background noise and account for its presence in their data. Rather than a starless sky, Min took measurements from parts of India he knew were not electrified and, using a similar process, distinguished the light that had been emitted from electricity from the other sources of light.

Bright Spot

The fruit of these efforts can now be seen on a website Min created in partnership with the World Bank, the NOAA, and Development Seed, a mapping and data visualization group. The site, Twenty Years of India Lights, allows anyone to find any village in India and track its light output over the past two decades. Min hopes the site will democratize the data.

The illuminated maps show where there are gaps in India’s electric systems. They can also help identify patterns in electricity access that are a result of political targeting or favoritism, a focal point of Min’s research. And they may inspire new ideas about how to use autonomously collected data to help the poor.

“We’re inundated with data, and the volume and speed of data collection is larger than it’s ever been before,” says Min, “but the big challenge, I think, isn’t a technical one. The much harder question is, how do we understand what we’re seeing?” 

• Learn About Supporting the Department of Political Science
   

Photo courtesy of NASA

Photo by Chris Hondros/Getty Images

Recalling his time at U-M, Patrick Ball (Ph.D. ’98) doesn’t name the usual places. Sure, he knows the Fishbowl and the Diag, but he took a few important detours while getting his sociology Ph.D. The first was to El Salvador, which, when he arrived in 1991, had already endured 11 years of civil war. There, Ball worked as a nonviolent accompanier by using his U.S. citizenship to protect activists who might be targets of political violence. But by the end of 1991, Ball decided the time for nonviolent accompaniment had passed. His colleagues asked if he knew anything about statistics and computers. “Oddly enough,” Ball recalls saying, “I did.

”Rather than words, Ball narrates human rights violations using numbers. The director of research at the Human Rights Data Analysis Group, an organization Ball started as a graduate student, Ball has testified in six war-crimes trials and nine official truth commissions, and he has advised the United Nations and dozens of non-governmental organizations.

“I’m not a war crimes investigator in the sense that I track someone like a police officer would,” he explains. “I’m looking for the pattern rather than the individual event.”

Safety in Numbers

Human rights violations are notoriously difficult to prosecute. Abuse is usually perpetrated in secret, and in a war people can die indiscriminately from battles and bombs as readily as they can by acts of genocide. Examining morgue records and exhumation reports and identifying an intent, such as systematic detention or ethnic cleansing, is tricky. The accused can counter that the source is biased, or the evidence is incomplete. They can argue that it’s one person’s story against another’s. Ball saw a different approach.

“You can’t only look at the data you have. You also need to account for what you don’t have,” Ball says. This new approach used quantitative analysis and reasoning to recognize patterns of violence and even to estimate events that the data don’t include. This new evidence was sound, impartial, and it stood up in court. It also created an opening for a new human rights narrative
to emerge.

The Numbers Don't Lie

When he begins an investigation, Ball usually doesn’t have much to go on.

The lists could be census data or handwritten notes of how many people crossed a border. They can include hundreds of thousands of complete and incomplete names as well as inaccuracies. They might say people are dead who aren’t or include people who don’t exist. Ball’s methods are good at filtering through these kinds of bad data—especially compared to surveys, which, because one person represents hundreds or thousands of people, can be inaccurate. To be sure, his methods stand up to bad data, and he tests them by intentionally adding errors to his analysis.

“You have to add a lot of garbage or delete a lot of stuff—more than 10 percent—before the results start to change,” he says. “And when you’re dealing with hundreds of thousands of original records, it becomes hard to imagine how you could successfully add invented information or deliberately falsify thousands of records without being detected.”

Ball’s results don’t prove that a particular event happened. His numbers can corroborate a working hypothesis, and they can prove that alternative explanations couldn’t have happened. For example, in the trial against the former President of Yugoslavia Slobodan Milošević, Ball graphed Kosovar killings to show they didn’t match up with NATO bombings or attacks from Albanian guerilla groups, and, therefore, couldn’t have been caused by NATO or the KLA. The killings and the Kosovar migration patterns did correspond, however, to the Yugoslavian government’s actions, evidence that was consistent with the prosecution’s argument that President Milošević was responsible for the violence.

The Devils in the Details

Accuracy and rigor are essential, says Ball, in order to get the truth. “We have to be right,” he says. “We have to be as accurate as possible, and that means knowing how to interrogate raw data’s silences. Otherwise your results become captive to those silences. You just reproduce them.”

In addition to Milošević, Ball testified against Guatemala’s General Jose Efraín Ríos Montt, who was convicted of genocide and crimes against humanity in 2013. He testified in the trial of former Chadian President Hissène Habré, who was convicted of war crimes and crimes against humanity in May 2016. Ball has currently trained his attention on Syria.

“Our Syrian partners stagger me with their persistence and meticulousness and courage,” Ball says. “I’m doing my best to honor and respect their work by doing the best record linkage I can.

“We don’t want to treat two different people as the same person,” he continues. “And that’s hard because a third of all Syrian men have the name Muhammad, and something like 85 percent of the victims in Syria are adult men. So this isn’t simple.”

But Ball believes doing this laborious, difficult work makes the world a more just place because it leads to accountability.

“My part is a specific, narrow piece, which just happens to fit with the skills I have,” he says. “I don’t think that what we do is in any way the best or most important part of human rights activism. Sometimes, we are just a footnote—but we are a really good footnote.” 

• Learn About Supporting the Department of Sociology

Photo by Martin Bruining/Getty Images

Among LSA junior Patrick Mullan-Koufopoulos’s earliest memories are the Saturdays he and his mother went to the jail near their home on the Greek island of Lesbos. They were there because his mother helped to run a non-governmental organization called—in English—Coexistence, one of the very few refugee organizations on Lesbos before the fall of 2015. As a small boy, Mullan-Koufopoulos didn’t understand these weekly visits until he threw a tantrum because a child inside a cell would not come out and play.

“My mom explained that he had to stay in that cell because he wasn’t free in Greece,” he recalls. “And I began to understand what it meant to be a refugee.”

Two and a half miles across the Aegean Sea from Turkey, Lesbos has long been a doorway to Europe. In 2015, Lesbos, an island of 86,000, found itself moving center stage in the drama of international displacement. That year, more than 900,000 refugees flooded into Greece from Turkey; on Lesbos, at the peak, that meant 3,300 refugees per day.

The scale of human suffering was overwhelming, and Mullan-Koufopoulos, flying back to Lesbos between semesters, met it the instant he stepped off the plane. “The airport is across the street from the ocean,” he says. “We went straight to the beach because a boat was coming in.”

In December, the sunny Greek islands are cold, windy, and rainy. The refugees were exhausted and soaked; many were crying. Volunteers waded into the water to guide the boats to shore. They helped the passengers, beginning with babies, who were passed along a line where they would be diapered, washed, and given some juice. Next were the children, then the next person, and the next. “That first day, it seemed so normal to everyone, which was shocking to me,” he says. “But within two weeks it became normal to me, too.”

Twenty boats might arrive in an hour, which meant helping people off one boat, and then moving to the next. The volunteers developed a system: first blankets, then tea, food, and dry clothes. Then the refugees boarded the United Nations High Commissioner for Refugees (UNHCR) bus to a detention center called the Moria, where they would be processed and either given papers to continue into Europe or turned away.

Lesbos is a way station, and a meaningful one. It’s a foothold to Europe that many refugees consider an end goal, though it’s also where refugees realize how much farther they have to go. It’s an extreme emotional space Mullan-Koufopoulos could enter because he spoke some Arabic and was a Greek American. “They thought Europeans and Americans were racist, that we all opposed immigrants and refugees,” he recalls. When they learned he was an American, many asked, “Why did you make ISIS?”

It was important to Mullan-Koufopoulos and his fellow volunteers that welcoming refugees not become rote. “On the front lines, safety was most important,” he says. “But humanity came next.” Every day, an elderly neighbor walked up and down the beach collecting the refugees’ wet clothes and shoes. He’d take them home, clean and dry them, and hand them out the next day. On Christmas, volunteers collected several of the thousands of life jackets that litter the island to create a Christmas tree. “It was a really nice way of taking away the religious connotation of the holiday to make it an experience for intercultural bonding.”

The refugees kept their individuality, too. “One woman said to my mother, ‘I used to be you,’” he recalls. “‘I had two kids, my husband was a professor. And now I’m here.’”

In March 2016, the European Union and Turkey struck a “one in, one out” agreement to return each refugee arriving in Greece to Turkey, and in exchange give one asylum seeker still in Turkey a European home. Many Greeks didn’t support the agreement, fearing it would encourage smugglers to send refugees on an even more dangerous route. A few weeks after the deal was struck, a boat carrying 500 refugees sank. “A lot of Greeks pointed at the EU,” says Mullan-Koufopoulos, “and said, this one’s on you.”

Mullan-Koufopoulos’s family is still working on behalf of the refugees. In Ann Arbor, he worries about them, but his belief that individuals can overcome institutional failure remains strong.

“At our best, we turned a system that had failed many people many times into a single situation where we didn’t fail them,” he says. “We created a place, for at least a few minutes, that offered them dignity and kindness in the purest way possible.” 

• Learn About Supporting the Residential College