First of all, being a data scientist can mean a lot of things. There are several roles that data people can fit themselves into. For instance, you could grow a narrow and specialized set of skills so that you would mostly craft (wonderful) visualizations, as would an artist, or only handle and maintain datasets as a data curator, or mainly design and deploy very complicated models as a core data scientist.
On the other hand, you can grow a very broad skillset and turn into a kind of full-stack data scientist. Each career path will request specific abilities and skills while coming with distinct challenges, rewards, and risks. Yes, risks. For example, a core data scientist in a small company may be replaced by H2O's Driverless AI; as ironic as it sounds...