Designing successful and cost-effective wildlife reserves presents many challenging network design problems. The goal of this work is to design robust and scalable algorithms for selecting reserve sites that together maximize ecologically important objectives such as the protected target species population size and connectivity within the reserve. I am specifically interested in algorithms for conservation decision-making under uncertainty and for sequential decision-making.
Invasive species both threaten biodiversity and cause billions of dollars in economic losses, necessitating effective strategies for controlling their spread. I work on developing new models for characterizing the spread of these species and the effect of targeted treatments in order to determine optimal intervention plans in budget-constrained settings.
Climate change poses a serious threat to critical infrastructure systems on which we depend. As participants in the 2017 UN Data for Climate Action Challenge, we developed a computational framework for determining how to allocate funds to improve the flood resilience of road infrastructure in Senegal.
A project using tensor factorization to determine whether there are distinct subtypes within the autism spectrum that are diagnostically significant.
A web-based interactive visual analytics tool for comparing bikeshare schemes across 10 US cities. The interface included back-end algorithms (ST-DBSCAN, Louvain modularity optimization and max clique detection) for detecting flow patterns and highly interconnected sets of stations.