There's a really rich set of analyses in http://blog.stephenwolfram.com/2012/03/the-personal-analytic... . "Times he likes to send emails" is a poor description for that body of work [I know the people who worked on it, and they worked hard and smart on it for a long time].
While I like neither buzzword, "data science" is probably better than "big data", for the following reasons:
1. not all data needs to be big to be interesting
2. the majority of 'science' that is typically done on data, both by enthusiasts and corporations, is both stereotyped and shallow. it won't be long before this is disrupted. unfortunately "big data" makes it sound like the problem is an engineering one -- in reality, the problem is cultural.
3. like 'traditional' science, data science is irreducibly hard. You need to be smart, creative, to know a diversity of methods, and you need interactive tools that allow you to explore and test hypotheses.
While I like neither buzzword, "data science" is probably better than "big data", for the following reasons:
1. not all data needs to be big to be interesting
2. the majority of 'science' that is typically done on data, both by enthusiasts and corporations, is both stereotyped and shallow. it won't be long before this is disrupted. unfortunately "big data" makes it sound like the problem is an engineering one -- in reality, the problem is cultural.
3. like 'traditional' science, data science is irreducibly hard. You need to be smart, creative, to know a diversity of methods, and you need interactive tools that allow you to explore and test hypotheses.