What is the role of the vascular scientist?

What is the relationship between the emergence of the "data scientist" role and NoSQL/Hadoop-based processing?

• Lots of stuff in the press about NoSQL data stores and Hadoop-based systems for data processing. Also lots of press about the emergence of the "data scientist" role, which is a lot more unstructured and algorithmic than classical business analyst/statistician roles. What's the relationship between the trends?  Why weren't there data scientists in the world of RDBMS's and Data Warehouse BI?  Was it because the scope of feasible analysis was more narrow previously?  Or are the "BI" tools of the NoSQL/Hadoop world just too immature?  Or do you believe even with greater maturity, there is a need for a data scientist in the Big Data world even though there didn't really seem to be one in the traditional data world?

• Answer:

  That is a loaded question. In a RDB it is pretty easy to draw connections, well bc they are mapped for you in a ERD. So the term scientist which implies more a hypothesis and testing role isn't really required if you're discovering any new relationships. The relationship is established already you're just manipulating it. With bigdata you have data from all sources in different units and there is no direct relationship drawn out for you. You have to create the relationship, and ten be able to prove it. You don't have to do a z-normalization if you're working in a RDB. So taking Twitter and Facebook data and evaluating that with Walmart Elmo doll sales is a little bit more involved than... Select elmo_sales from sales_data where sales_dt > 'dec-12-2012' The scientific part is not regarding the extraction of data it is the discovery of relationships that makes it a science. In my opinion...