In this talk, I'll describe what I see the three pillars of data
science are, and specifically how data scientists often fail at one of
these pillars: causal inference. I'll provide three examples of where
understanding causal inference has been necessary.
The first example is a classic causal inference result on a dataset
involving infant mortality and birth weight. The example more
generally applies as a cautionary lesson to anyone doing analytics.
The second example is a showcase of causal inference being used at
Shopify to answer million dollar questions about shopping behaviour.
Finally, the third example is an application to fairness and
unbiasedness in production machine learning systems.
Cameron Davidson-Pilon has worked in many areas of applied statistics, from the evolutionary dynamics of genes to modelling of financial prices. His contributions to the community include lifelines, an implementation of survival analysis in Python,lifetimes, and Bayesian Methods for Hackers, an open source book & printed book on Bayesian analysis. His previous education was at Wilfred Laurier University, University of Waterloo and the Independent University of Moscow, and he currently works as a Director of Data Science at Shopify in Ottawa, Ontario.
Contact at the MS2Discovery Research Institute: Manuele Santoprete
(Host of the speaker, Multidisciplinary Talk, Tectons 2,6,7 and
Refreshments will be provided