Anthony M. Townsend - Smart Cities - Big Data, Civic Hackers, and the Quest for a New Utopia

Most of the world’s cities are now lit up by some kind of wireless service. But as Ericsson, a leading supplier of network equipment, points out, “Reaching the next billion subscribers means expanding to rural off-grid areas.”24 The company has developed highly efficient solar-powered cell towers for use in outlying areas where there is no electric-power infrastructure. On the consumer side, in 2010 Vodafone launched a $32 solar-powered phone in India. Presumably, the arrival of modern telecommunications in the countryside might provide new local economic opportunities and slow migration to cities. But it could just as likely accelerate migration by plugging ever-larger rural areas into the social and economic life of the city. One study that tracked migration through mobile phones in Kenya uncovered an astonishingly high turnover rate for new arrivals—on average, newcomers during a year-long study period in 2008-2009 stayed in Kibera, the capital’s largest slum, just less than two months. Anthropologist Mirjam de Bruijn has documented Bedouin caravans in the southern Sahara that have altered their historic trade routes to periodically pass through areas of mobile phone service. Even indigenous peoples want to stay connected in a global economy.

Development organizations are just beginning to wrap their thinking around the tremendous opportunity for development that mobile phones present. Richard Heeks, the professor of development informatics, sees a marked shift in the ICT4D movement from PCs to mobile devices. “We stand at a fork in the Internet access road,” he wrote at the conclusion of his 2008 article. “We can keep pushing down the PC-based route when less than 0.5 percent of African villages so far have a link this way. Or we can jump ship to a technology that has already reached many poor communities.” It isn’t just scholars and activists calling for a new model. In January 2013, when Google chairman Eric Schmidt spent a week visiting a handful of booming African cities, he saw firsthand the role of technology as a tool for economic opportunity. “This new generation expects more, and will use mobile computing to get it,” he reported.

Over the next decade, mobiles promise to become even cheaper and more pervasive. Assuming even a modest rate of replacement and a continued drop in smartphone prices, it is very likely that a decade from now half of the world’s people—including hundreds of millions of the urban poor—will be walking around with devices that are essentially supercomputers in their pockets. Broadband wireless networks with data speeds in excess of 100 megabits per second or more will light up entire cities, including their slums.

But mobiles aren’t simply new economic tools for the world’s urban poor. Increasingly, mobile networks themselves are becoming observatories where we can watch in real time how people move, how cities grow, the quality of life, and economic activity.

Taking the Global Pulse

The lights went down on a room full of diplomats at the United Nation’s General Assembly in New York in November 2011. “Imagine it’s 2009, the rains are late, and food and fuel prices are rising,” said Robert Kirkpatrick, director of the organization’s Global Pulse project. “What would it have looked like in data collected by a mobile operator?”30 He rattled off a list of telltale signs of distress. People might shift to smaller, more frequent purchases of airtime as their economic anxiety increased. Increased defaults on microloans would show up in payment systems like M-Pesa. Calls to livestock dealers would spike as families liquidated agricultural assets to survive. Phones purchased in villages would suddenly request connections with urban cell towers, as displaced farmers flooded the city looking for work.

The financial crisis of 2008 hit the world’s poor hard. Food and fuel prices were already rising just as the contagion spreading through global financial markets released a parallel shock wave at the bottom of the pyramid. As Secretary-General Ban Ki-moon explained at the same event, UN officials were certain that the crisis would “inflict suffering immediately on the poorest and most vulnerable” people. But the economic chain reaction moved faster than his statisticians could track it. “It was clear we were seeing something new,” he continued, “Impacts of the crisis were flowing across borders at alarming velocity.” A decades worth of economic gains evaporated overnight as hundreds of millions of families slipped back into poverty.

Ban Ki-moon moved quickly and decisively (by UN standards). “Our need for policy agility has never been greater,” he explained, “Our traditional twentieth- century tools for tracking international development can’t keep up. By the time we can measure what’s happening at the household level, the harm has already been done.” With the governments of the UK and Sweden serving as angel investors, Global Pulse was launched in April 2009 as the Global Impact Vulnerability Alert System (it was later re-branded). It was charged with developing new sources of real-time data to create an early warning system for social and economic crises.

Global Pulse promised to be the biggest advance in public demography in a generation. Consider the venerable US census, which is far more thorough and consistent than what poor countries can do. It swallows up huge amounts of resources and routinely misses millions of people. Since it is done only once a decade, there have only been a few dozen chances to improve it. While other less comprehensive interim surveys are taken to update the results, the master count happens just once per decade because it requires an army of over six hundred thousand census takers to collect data house by house. The United Nation’s own methods are similarly plodding. As Global Pulse’s 2011 annual report observed, “Traditional data collection methods like door-to-door household surveys ... can take months or even years to complete and are woefully inadequate for this task.”

At the opposite end of the spectrum are the tools used by market researchers and pollsters. Free from the constraints that hamper government data collection, they can collect information almost anytime using any survey and statistical method at hand. They can tweak survey questions day-to-day to home in on emerging trends and fine-tune their observations. And they can go beyond surveys and tap nearly limitless pools of real-time private data on credit-card transactions, store visits, or web-browsing habits. Instead of sifting through the tailings of macroeconomic statistics for clues about recent events, they can plug into a sensory infrastructure that shows what is happening in the real economy at a microscopic level, second by second.

To bring the United Nation’s crisis-sensing abilities up to date, Kirkpatrick partnered with a variety of research partners around the world to explore new ways of picking up signs of distress in the social and economic data exhaust of poor nations. One of the most promising experiments was done with Jana, a Boston-based company that had developed a tool for conducting surveys by mobile phone. Jana was the brainchild of MIT Media Lab alum Nathan Eagle, who spent several years in Kenya teaching students how to develop mobile phone apps. While working on a tool for nurses to report on blood supplies at rural clinics by text message, he noticed that participation quickly fell off. He needed a way to reward the nurses for responding to the text messages asking for updates on blood inventory. After returning to the United States, Eagle developed a system for compensating survey respondents with tiny amounts of airtime. Jana now has partnerships with hundreds of mobile phone companies, and can reach over 2 billion potential respondents worldwide. Global Pulse put Jana’s system to work by sending out short queries by SMS, such as “Were you sick in the past 7 days?” or “If you had 15 USD what would you spend it on?” The thousands of respondents were rewarded with free airtime from Jana’s servers, which are wired directly into carriers’ billing systems. Kirkpatrick insisted that these guerrilla surveys wouldn’t replace traditional data collection efforts, but were intended rather to plug gaps and help inform the design of more traditional surveys. But if this approach proves accurate and reliable enough for day-to-day use, the data-gathering capabilities of poor nations could quickly leapfrog that of rich ones.