My take... Historically there are specialists for data such as Business Intelligence, Network Architecture, ETL, etc. They know how to move, manage and scale data.
Then there are the analysts that use the data such as Statisticians, Engineers, Physicists, etc. This is more the applied side of data.
The advent of computing power and speed of networks has allowed data to be more accessible. Now there is a need to merge these two specialities together. The Data Scientist knows enough of ETL and knows enough of applied math and statistics. This may be more of a generalist role.
> On the term “data scientist”, Silver once said that a “data scientist is a sexed up term for a statistician. Statistics is a branch of science. Data scientist is slightly redundant in some way and people shouldn’t berate the term statistician”
Nate Silver is wrong. Statistics is not a “branch of science”, and being a data scientist, in practice, is not quite the same as being a statistician. Data scientists use statistics as a tool (just like regular scientists, such as physicists, do), but they are not statisticians. Being a data scientist is not that different from being a regular scientist, but, instead of studying some physical phenomena for fundamental purposes (to learn more about the Universe or our bodies), a data scientist studies some very specific set of phenomena relevant to his industry and builds predictive models that have business value. Statistics is but one (actually small) part of.
Statistics is not part of science to the exact same extent that maths is not part of science: by a technicality at most. You also have a very narrow view of what a statistician is if you think they don't spend most of their time applying methods on real data.
I would say that being a "data scientist" is quite different from being a "scientist", with "data science" being more a kind of engineering than a science. For example, the goals of the engineer and the data scientist coincide in the sense that both seek to build products for business purposes. Further, their methods are of a similar kind as in their work towards the product they use statistical models created in house or the industry at large that exist in conceptual frameworks from academia.
I don't necessarily disagree with you. But there is a distinction between what you call 'regular scientist' and 'data scientists' right? Pretty much all 'regular scientists' have PhDs in their discipline and are working on cutting edge things in their field of expertise. The term 'Data Scientist' is used much more loosely. Besides, do a lot of those regular scientists call themselves scientists? The term 'scientist' has an air to it. To some, 'data scientist' sounds more presumptuous that 'machine learning engineer' or 'senior modeler' or 'statistician'
I can see why the term 'statistician' doesn't seem to cover everything in a data scientist's skill set. Perhaps we need an alternative moniker.