DATA SCIENCE IN THE SERVICE OF PALEONTOLOGY
As an industrial statistician (or data scientist in the modern vernacular) and amateur fossil collector, I describe a few projects where I was able to contribute quantitative methods that produced publishable research. By being a separate set of eyes without the pull to orthodox approaches in paleoecology, new insights were obtained. These collaborative efforts enriched my industrial career as well, by forcing me to learn new statistical methods and to develop consulting skills for a field that is not quantitative at its core. For the most part, I was able to apply and develop “dual purpose” approaches, resulting in synergy between my industrial and amateur-paleontologist lives.
I will briefly describe detecting coordinated stasis in the Devonian fossils of the Appalachian Basin of New York, the theory that asserts community structures remain relatively constant over millions of years. The method used here has its roots in text mining. I will also describe how to use a statistical model of exposure risk to identify predation as a likely factor in mollusk ornamentation evolution. Bayesian methods were essential to the success of this project.
This entire thread of my professional life began with my membership in the Rochester Academy of Science and its Fossil Section. I was introduced to academic researchers and their work through our annual poster session and through the many talks our Section hosts each year. The Section also allowed me to serve the public through being a club officer and contributing to our outreach events.