“Contrary to common belief, the hardest part of data science isn’t building an accurate model or obtaining good, clean data. It is much harder to define feasible problems and come up with reasonable ways of measuring solutions.” Yanir Seroussi ( 23.11.2015 )