Spatially Defined Measures of Mobility from Call Data Records: Developing Big Data Measurements and Applications for Social Science

Censuses and surveys have been the primary source of information on mobility and migration. However, concerns with these data include sample size, detail, accuracy and expense. In the past decade, large scale mobile phone data has recently become available for the study of human movement patterns and holds immense promise for studying human behavior on a vast scale never before possible and with a precision and accuracy never before possible with surveys or other data collection techniques. There is already a significant body of literature that has made key inroads into understanding mobility using this exciting new data source, and there have been several different measures of mobility used. However, there has been little discussion and analysis of these measures. It is unclear what exactly these measures measure and we argue that existing measures are contaminated by infrastructure and demographic and social characteristics of a population. These issues would be best addressed immediately as they will influence future studies of mobility using mobile phone data. In this presentation, we will describe new methods for measuring mobility with mobile phone data that address these concerns. Our measures are designed to address the spatial and social nature of human mobility. The practical relevance of massive geolocated data in the context of HIV research will also be discussed. Joint work with Nathalie Williams, Tim Thomas and Matt Dunbar.

Speaker Details

Adrian Dobra is an Associate Professor of Statistics and Nursing with a joint appointment in the Department of Statistics and the Department of Biobehavioral Nursing and Health Systems, University of Washington. He is also a core member of the Center for Statistics and the Social Sciences (CSSS). He has made methodological contributions to Bayesian statistics and to categorical data analysis with a particular focus on the development of multi-way Gaussian graphical models of various kinds (Gaussian graphical models, graphical log-linear models, dependency networks, Bayesian networks). These models can be applied in a Bayesian hierarchical models framework to capture complex patterns of multivariate dependence, including spatial associations. His work has also included the analysis of large-scale gene expression patterns. His areas of applied research are related to the social sciences, spatial epidemiology, genomics and disclosure limitation. Most recently he became involved in human mobility research based on geo-located temporal data recorded by wearable sensors.