Data forensics is the mining and analysis of data (data analytics) to extract dependencies and correlations between events and infer causes for undesirable symptoms in computer systems and the physical world.
Data forensics allows system administrators to keep their machines running as smoothly as possible—and to fix them when they don’t. Kai Shen, associate professor of computer science, is one of the leaders in the field. His expertise is in understanding the causes of negative performance anomalies, such as sudden reductions in throughput or responsiveness.
Performance can plummet for unexpected reasons: users may find they can’t load web pages, sound or video output may “break up,” and energy consumption or heat generation may spike to dangerous levels. Users may be particularly frustrated because “everything was fine” a moment before. Shen uses data science to explore the causes of such nonlinear behavior and develops techniques that can be used to diagnose and fix them on the fly, thereby maintaining the integrity of systems we count on for business, finance, human resources, customer service, manufacturing, and beyond.