“This podcast is a very informal and somewhat rambling discussion around applied statistics and data science. It takes the form of a conversation between Hilary Parker, a data scientist at Etsy, and Roger Peng, a biomedical statistician at Johns Hopkins (author of various Leanpub books and Coursera courses).
The podcast is pretty approachable, and the central theme seems to be comparing the perspectives of an an applied statistician in industry with that of a tenured academic. Topics include theory vs. practice, non-statisticians (such as devops people) tackling statistical problems, and issues around practicality, correctness and reproducibility.
Roger and Hilary both come at data science from the statistical perspective and both use R day-to-day, but also talk around other approaches and tool-chains, such as machine learning, the Python data science stack etc.
There’s a lot of humour and practical advice mixed in, and while the discussions include many casual statistical references (e.g. confounders, variance, test statistics, t-tests, skew, long-tail, Hadley, Hadleyverse, dplyr, ggplot etc.), I suspect listeners with no background in stats could just let those terms wash over them and still enjoy the discussions.
Recommended for anyone with an interest in practical data analysis.”
njr0 via Apple Podcasts ·
Great Britain ·
01/13/16