There is an easy and time tested solution for generating usable, reliable, and sometimes informative data. It’s called the scientific method, a key element of which is the design and execution of experiments. As scientists I would think that data scientists would be well schooled in the use of the scientific method and thus would be highly skilled in DOE and at least familiar if not experts in a wide variety of experimental methods from a number of scientific disciplines. After all this is what scientists do. Oh wait, sorry I forgot. Data ‘scientists’ are not scientists at all. Oh well, at least you get paid the big bucks.
This is a small problem today but it is poised to become much larger. Business and society cannot thrive when everyone analyzes data but nobody knows how to generate it as I wrote about recently.