Professor Ronald R. Coifman is a computational Harmonic Analyst developing method for data driven feature and and geometry extraction. Currently leading a program to derive scientific models from unlabeled observations. The goal is to define the underlying space/time /latent variables and corresponding model dynamics. These tools are applied to Neurology, Pathology, Medical data bases, and nonlinear Physics phenomena.

What do you do with data science?

Modeling unlabeled observational data to discover lintrinsic latent variables enable scientific analysis of the data, are tools are mostly geometric analytic, such as diffusion geometries
the goal is to develop the science behind the data.