Looking at big data to find patterns and minimize human conflict

Understanding how wars break out relies on understanding social divisions, which may be illuminated by data analysis.
Understanding how wars break out relies on understanding social divisions, which may be illuminated by data analysis.

When it comes to managing conflict, the human race has, to put it mildly, a spotty record. No matter our myriad similarities, society has always managed to split people along the same lines — religious, ethnic, economic, linguistic — since the dawn of civilization. The patterns of violence and conflict just keep repeating, with no solutions or understanding in sight.

Could today’s technology be tapped to end this cycle? More specifically, can big data reveal new ways to understand human behavior and the patterns that create conflict?

David Meyer believes it can. He’s a professor of mathematics at UC San Diego and a research scholar with the University of California’s Institute for Global Conflict and Cooperation, a multi-campus committee devoted to conflict research and the promotion of international cooperation.

“There is a long tradition of mathematical analysis of ‘big’ data on conflict, dating back at least to Lewis Fry Richardson’s work that began in the early 1900s,” Meyer said.


Lewis Fry Richardson was an English mathematician, physicist and meteorologist whose Quaker pacifism informed his work. He believed, as does Meyer, that conflict must be understood before it can be reduced. Richardson applied the mathematics he used to forecast weather to the study of relations between countries and looked for patterns in outbreaks of conflict.

“As he put it, ‘There is in the world a great deal of brilliant, witty political discussion which leads to no settled convictions [about the causes of war],’” Meyer said. “The data to which we now have access is immensely bigger than that data he collected. My hope is that our conclusions may be correspondingly more precise.

“I’ve been thinking about mathematical aspects of political conflict for a couple of years. Civil war results from [political conflict]. Understanding how civil wars spread relies on understanding social divisions.”

Harvests and violence


A collection of reports of violence published in regional Indonesian newspapers from 1990 through 2003, assembled by the United Nations, was among the first sets of “big data” Meyer examined. Mathematical analysis revealed cycles of violence on Java, the most densely populated Indonesian island, that peaked at four-month intervals in late April, late August and late December.

“The pattern was admittedly simple to discover but, at the time, completely puzzling,” Meyer said. Searching for an underlying cause, Meyer found that the peaks fell near the ends of rice-harvesting cycles.

Further digging revealed reports of violence among agricultural workers as far back as 1891. Meyer found that improvements in rice varieties have led to changes in how the crop is harvested, with fewer workers needed. Limited opportunities for both local villagers and migrant workers to participate and share in the harvest have led to fighting among the workers, landowners and foremen.

Meyer described the cyclical violence his team’s mathematical analysis had revealed and his ideas about the underlying causes at the annual meeting of the Association of Asian Studies in March.

Making connections

Meyer and his multidisciplinary team of UC San Diego data experts and social scientists recently set their sights on analyzing telecommunications data to identify and locate religious and ethnic groups in regions where census data is unavailable or out-of-date.

“This describes, for example, Iraq, where the last census was taken in 1987,” Meyer said. “There are important political reasons to know how many Kurds, Sunnis and Shias there are in Iraq, and where they live.”

A recent study of West African nation Ivory Coast’s calling patterns could pave the way. Meyer’s team for the project consisted of research scientists, postdoctoral scholars and students in mathematics, physics, computer engineering and political science at UC San Diego. Looking at language as a common thread, they analyzed a huge amount of cellphone data from Ivory Coast, where more than 70 regional languages are spoken.


Using the data, which was provided by French multinational telecommunications giant Orange, the team mapped connections among 1,216 cell towers between December 2011 and April 2012. The patterns and volume of calls passed between towers revealed communities defined by shared languages. A detailed map defined the communities, and in 2013 the project won a Data for Development Challenge award from Orange.

Such an approach could be useful in finding solutions to the volatile political and ethnic landscape in Iraq, a persistent roadblock to social justice and a unified government, Meyer said.

“Our results from Ivory Coast suggest that telecommunications data could be analyzed to estimate these demographics,” Meyer said. “Knowing them would provide some insight into possible future evolution of both political and violent conflict between these groups, and might suggest possible solutions.”

Robert Young, Brand Publishing Writer