My research utilizes computational social science and artificial intelligence to derive contextually informed algorithmic frameworks for understanding individuals and the social systems which influence their behavior, and for supporting positive behavior change within these systems. In particular, I analyze policy and behavior, and apply interactive behavior support and predictive algorithms, within the areas of higher education, environmental sustainability, and political participation.
Individuals are increasing spend their time in quasi-digital environments generating a wealth of data traces with which to understand their behavior. However, this data can produce its own set of challenges (sparsity, heterogeneity, connectivity) and generating new understandings of human behavior necessitates novel statistical and computational approaches. In my work I address these challenges by developing probabilistic frameworks which can capitalize on structure between data instances, producing state-of-the-art performance in tasks from predicting household appliance energy consumption with an energy disaggregation framework to predicting the movements of human traffickers.