Talk given to Statistical Society of Australa - Victorian Branch 2022-07-21.
Really Useful Engines: Engineering Data Science
The modern Data Science team dreams of impact, or being declared "Really Useful". The path to this goal is fraught with challenges, and in response we have become increasingly dependent on workflows steeped in code, fortified by Software Engineering principles. Slowly but surely it's become apparent to me that this fortification is insufficient. Data Science has unique failure modes, on which Software Engineering tools are mute. There remains plenty of scope for Data Science teams to become mired in "Confusion and Delay". In this talk I build on previous expositions of individual Data Scientist workflows in an attempt to distil principles that can transform Data Science teams into Data Science engines.
https://docs.google.com/presentation/d/1pzeLJV7I8xaYVWFsYOe0th811Xg9unjxCRiI_cP9v2E/edit?usp=sharing