Accurate and timely estimates of population characteristics are a critical input to research and policy, but reliable data is often scarce in developing and conflict-affected regions. In recent work, we have shown how machine learning algorithms can be applied to mobile phone metadata to infer fixed characteristics of individual subscribers, such as wealth and gender. Here, we describe efforts to extend this approach to a non-stationary regime, to detect and measure changes in an individual's welfare over time. For this study, we tracked 1200 Afghan citizens with high-frequency panel surveys, and matched each person's responses to psuedonymized transactions log of mobile phone activity. Preliminary results indicate that it is possible to detect negative shocks (e.g., violence), and positive shocks (e.g., receiving a gift) from the mobile phone records alone. This suggests the possibility of real-time tracking of vulnerability, and new paradigms for program monitoring and impact evaluation.
Joshua Blumenstock is an Assistant Professor at the U.C. Berkeley School of Information. His research develops theory and methods for the analysis of large-scale behavioral data, with a focus on how such data can be used to better understand poverty and economic development. Recent projects combine field experiments with big spatiotemporal network data to model decision-making in poor and conflict-affected regions of the world. Prior to joining UW, Joshua was on the faculty at the University of Washington, where he founded and co-directed the Data Science and Analytics Lab. He has a Ph.D. in Information Science and a M.A. in Economics from U.C. Berkeley, and Bachelor’s degrees in Computer Science and Physics from Wesleyan University. He is a recipient of the Intel Faculty Early Career Honor, a Gates Millennium Grand Challenge award, a Google Faculty Research Award, and a former fellow of the Thomas J. Watson Foundation and the Harvard Institutes of Medicine.