Absolutely. You have to marry human instinct with big data (or any other kind of data for that matter, I suppose). The flip side though sometimes is that human instinct is frequently wedded to traditional uses of data and sometimes big data challenges traditional assumptions (regarding the role of theory, for instance; or even something as basic as data quality). The 'training' may thus have to encourage people to question a few established assumptions about data... ...the assumptions may turn out to be well founded, but... Thanks so much for a terrific post.