A growing number of scientific fields suffer from a stifling embarassment of riches: data pile up much faster than they can be analyzed. A team of researchers at the University of Southern California's Information Sciences Institute is now building a prototype of a system that will address the problem by automating scientific workflows.