Data Science Research Group: Dr. John Handley

The Role of Data Science in PaleoecologyAbstract:Paleoecology is the study of past ecosystems with the aim to reconstruct ancient worlds and to understand the patterns and processes of evolution. It is fundamentally an observational science, based primarily on the fossil record and geochemical signals. Data collection is often restricted to available exposure of fossil-bearing rocks. These data are often augmented with museum collections where the sampling protocol is unknown. In this talk, I will describe several studies where sampling issues were dealt with using statistical modeling approaches. I will show examples of how multi-level models and Bayesian estimation can capture some of uncertainties associated with heterogeneous samples. Surprisingly, these methods are novel to paleoecologists and I will discuss some of the cultural barriers I overcame when working with scientists more comfortable with descriptive statistics than inference. Speaker Bio:John Handley is a research scientist who focuses on applying advanced statistical and data analytic methods in diverse fields such as image processing, psychometrics, urban analytics and mobility, service pricing and contracts, and quantitative paleoecology. In his role at the Rochester Data Science Consortium, he collaborates with industrial and academic scientists and engineers to drive economic impact. Prior to joining University of Rochester, John was a principal scientist at Xerox, Xerox PARC and Conduent, Inc. where he led a variety of projects in transportation, device management, and service contract analytics. He also served as a consultant to Bellcore and Lexis/Nexis in document image processing.  John holds B.S. and M.S. degrees in Mathematics from Ohio State University and received a Ph.D. in Imaging Science from Rochester Institute of Technology. He serves as a Research Associate with the Paleontological Research Institution affiliated with Cornell University.