The message is a good one unfortunately (through no fault of his own) the messenger lacks credibility since his own ‘science’ is not in fact science but is instead only a tool of science. A lecture on how to do science from a person claiming to be a scientist, but who is not actually one does not sit very well with real working scientists who, in addition to being actual scientists, do (some parts of) data ‘science’ every day as a part of their normal mode of doing science and being actual scientists. Until data science comes to terms with itself and recognizes itself for what it is save the lectures on doing science for the classroom or your fellow data analytics professionals. You don’t see research scientists lecturing data ‘scientists’ on how machines can’t actually learn do you? Oh, wait a minute..hmmm, well maybe that was a bad example. lol! Let’s try a different example. How about this? (Even though many could) you don’t see research scientists lecturing data ‘scientists’ on the appropriate use of regression, clustering, or factor analysis do you? You should take much solace in the fact that data ‘scientists’ make (way, way) more money than actual scientists so they are clearly valued at a higher level by society at the moment. It actually makes a lot of sense when you think about it. What’s so great about contributing to an understanding of the fundamental nature of the universe when you can figure out what voice belongs to a given celebrity instead? Actually, don’t think about it. It will make you very, very sad.