I know and work with people who would be considered "true" data scientists as well as some aspiring data scientists. The consensus seems to be, based on what I've read and among people that I've talked to with post graduate degrees in things like Predictive Analytics is that a Data Scientist:
1) Has a post graduate degree in Math, Statistics, Economics, Finance or some other math-heavy field of study
2) Has solid expertise in BI or Big Data... E.g. the MS BI Stack or AWS. They may be a highly certified or just have several years of solid experience
3) Has expertise in a specific business subject area... Banking, Finance, Advertising, etc...
Such a person could understands a specific business, software and has the math skills to develop a solution that helps them turn mad amounts of data into actionable intelligence. That's my understanding of what a data scientist is and what they do. I'm a BI who works in big data but don't see myself going back to school to become a data scientist.
What I've seen is, because the person I just described is as rare as a unicorn many companies will opt for a data science team that consists of an analytics person(s), a BI person(s) and visual experience person(s) (e.g. a Tableau guru) and they get their expertise about the business from one or more people who are SME's in the business about that business.
Again, this is based on my conversations with people and what I've read. I'm interested in seeing how Big Data and data science evolves.
"I cant stress enough the importance of switching from a sequential files mindset to set-based thinking. After you make the switch, you can spend your time tuning and optimizing your queries instead of maintaining lengthy, poor-performing code."
-- Itzik Ben-Gan 2001