Presentation: Data Science as a Science—Large Scale Online Experiments in How People Do Data Science
2018 Student-Invited Speaker: Jeffrey Leek, Ph.D., Professor of Biostatistics, Johns Hopkins Bloomberg School of Public Health
Abstract: There have been decades of work focused on the mathematical and theoretical optimality properties of methods and software for data science. But this theory has come crashing up against a problem that math can't solve - human behavior. In this talk I will make the argument that fake news, the reproducibility crisis, and the arguments over p-values and p-hacking can all be attributed to our lack of understanding about human-data interaction. I will then discuss some observational studies and experiments we have been performing to study the ways that humans interact with data. This builds on existing work in the data visualization space, but expands these experiments to consider the entire data science process.