I stumbled upon the Data Skeptic podcast about a week ago, and so far it has been amazing. I’ve consumed podcasts for years, in various categories, and I can happily say that this is one of the best podcasts I’ve ever subscribed to.
The host is a live and breathe data scientist – apparently he had Bayesian beliefs incorporated into his wedding vows – and he takes to the microphone very well. This combination of knowledge and passion for the subject matter, a voice made for radio, and what seems like really careful scripting makes for a killer combination.
Most shows, called ‘mini episodes’, are about 15 minutes long. In that time the host explains some data science technique to his wife, who seems to be a layperson. He doesn’t dumb it down or stick to abstract or frivolous concepts; he’s described ant colony optimisation, bayesian reasoning, cross validation, and other “meaty” concepts and techniques. Other shows are full length (about 1 hour long) and feature guests who are experts in some industry or application domain.
I will ditch a podcast due to poor production value, poor flow, or narration simply being unpleasant (I want to like Data Stories, but I find most episodes quite difficult to listen to). The Data Skeptic seems to be in little danger of any of those potential pitfalls, with one possible caution: the expert interviews are a bit on the long side, and if the quality of guests (not in term of knowledge/expertise, but in terms of how well they take to the podcast/radio format) can vary. It might be a good idea to shorten the full length interviews.
Really glad I found this podcast, I’m loving it, and the mini episodes are especially valuable. Anyone with even a slight interest in data science should subscribe.