Investigation of Complex Social Systems from Communication Data: Scaling from 102 to 108

Speaker  Nathan Eagle

Host  Eric Horvitz

Affiliation  Santa Fe Institute and MIT

Duration  01:17:59

Date recorded  29 June 2009

Records of human movements and communication patterns are continuously being generated by the 4 billion mobile phone users around the world. In collaboration with mobile operators in dozens of countries, my colleagues and I are aggregating and analyzing this behavioral data from over 250 million people from North and South America, Europe, and Africa. This unprecedented volume of information about movement, purchasing decisions, and social networks has led to a novel set of research questions applicable to a wide range of academic disciplines. This data enables comparisons across cultures and continents, and may even lead to the identification of universal laws governing human behavior. I will discuss a selection of results related to the scaling of complex social systems; ranging from inferences about human relationships on a social network of 100 people, to the evolution of community structure on a national communication graph of 1 million people, to topological properties correlated with socioeconomic status on a communication graph representing over 100 million people.

I believe the ultimate value of this research will be in its application to solving global social challenges. We are now in a position to incorporate insights gained from these data into the design and engineering of social systems in order to actively improve human lives; we have named this nascent field Engineering-Social Systems. Understanding the behavioral patterns of millions of people is of more than academic interest, and we believe the ultimate question is the engineering one: how can we put those insights to good use? To that end, I will provide a broad overview of upcoming projects using behavioral data related to public policy, urban planning, developmental economics, and computational epidemiology.

©2009 Microsoft Corporation. All rights reserved.
> Investigation of Complex Social Systems from Communication Data: Scaling from 10<sup>2</sup> to 10<sup>8</sup>