There are over 4 billion mobile phone subscribers worldwide, the majority of whom live in the developing world; their phones are continuously generating vast volumes of movement, communication, and even financial transaction data. This rapid global penetration of the mobile phone enables novel approaches to the design and implementation of public health research. The ESS Lab aims to explore how this newly ubiquitous mobile phone infrastructure and unprecedented volume of behavioral data can be repurposed for global public health from the individual to the population scale. At the level of the individual, we have provided subjects, ranging from male sex workers in Kenya to smokers in New York, with special smartphones programmed to send continuously-recorded behavioral data directly to our research team. On a population level, the ability to send surveys, information, and financial compensation to the majority of a nation provides unprecedented new scope for gathering longitudinal health-related data. On a global level, we have collected up to five years’ worth of movement and communication data from 350 million mobile phone subscribers around the world – one of the largest and most diverse human behavioral datasets ever studied.
We are using this data in a variety of projects including studying the efficacy of Mexico’s containment strategies during the recent H1N1 outbreak, and designing algorithms that mine this data to automatically detect cholera outbreaks. These projects will require the parallel development of a suite of new computational tools and machine learning algorithms capable of petabyte-scale data analysis, which we believe will ultimately lead to major advancements in global public health.
Catastrophe Modeling for Rwandan Disease Surveillance
Can mobile phones be used as an early warning system for disease outbreaks? Bayesian anomaly detection algorithms may be able to quantify behavioral signatures associated with cholera outbreaks in Rwanda. If successful, these alogithms could lead to the deployment of next generation of disease surveillance systems in some of the world's regions that need it the most. - A. Kapoor, N. Eagle, E. Horvitz
Spatiotemporal Diffusion of Contraceptive Norms in the D.R.
How do contraceptive norms spread through rural areas of the developing world? Spatiotemporal diffusion models have the potential to better evaluate the efficacy of HIV prevention techniques and inform policy decisions related to public health. - H. Yoshioka, N. Eagle
A Causal Model for Quality of Schooling
A key challenge for policymakers in many developing countries is to decide which intervention or collection of interventions works best to improve learning outcomes in their schools. Our aim is to develop a causal model that explains student learning outcomes in terms of observable characteristics as well as conditions and processes difficult to observe directly. - N. McGinn, M. Moussavi
Mobility Models of Malaria in East Africa
How do human mobility patterns affect the spread of malaria? Aggregating longitudinal movement data from 15M mobile phones in East Africa, it may be possible to gain a better understanding of the implications of human movement on the spread of disease. - A. Wesolowski, N. Eagle, A. Tatem, D. Smith, A. Noor, R. Snow, C. Buckee