ABSTRACT
An enormous amount of behavioral data is now being generated and stored by billions of individuals across countries and cultures. e ubiquity of mobile sensors from laptops to cellular phones has ushered in a new age of digital data storage and management. It is now possible to measure social systems on a massive scale. Mobility patterns of city dwellers can be inferred from mobile phone use [4], epidemics can be
CONTENTS 11.1 Introduction 331 11.2 Data 333
11.2.1 Conditioning the Data 335 11.3 Methods 338
11.3.1 Basic Analysis and Statistics 339 11.3.2 Auto-and Cross-Correlation 343 11.3.3 Correlation Matrices 345 11.3.4 e Eigenvalue Spectrum and Comparison to Random
Matrices 347 11.3.5 Daily Drug-Related Crime Rates in 1999 348 11.3.6 Weekly e- Related Crime Rates from 1991 to 1999 350
11.4 Summary and Conclusion 357 Acknowledgments 358 References 358
modeled across distances of many orders of magnitude [20], and the diffusion of information can be measured for large populations [10]. Armed with this ood of data, there are very real and important opportunities to study and ultimately facilitate sustainability in human-built systems. Insight into these systems can help inform the social sciences from economics to sociology as well as provide policy makers with critical answers that may be used to better allocate scarce resources or implement bene- cial social programs.
