PyData Global 2022 talk on reliable data science

Last Thursday, I gave a talk at PyData Global 2022. In many ways, the talk picks up right where we left off in the SciPy 2016 talk, but adds many more lessons learned from the intervening years of consulting and running the DrivenData challenge platform since then.

One of the fun things prepping the talk was (loosely) connecting challenges in data science to a few interesting concepts about safety and risk mitigation that I was exposed to in the Coast Guard:


Previous: Bayesian equivalent of power analysis

Any comments or suggestions? Let me know.