Depends on your familiarity with the unix toolset vs use cases, an amusing (if somewhat contrived) example of this:

http://aadrake.com/command-line-tools-can-be-235x-faster-tha...

Most CSV analysis is static, not something ongoing with inserts/updates/etc, so using a combination of unix tools to handle the data processing seems like a sufficient solution for many scenarios.

I use CSV all the time because I am constantly generating new datasets and it is nice to be able to literally cat|pbcopy them and paste them into DataGraph to see the results. Every major language has some sort of csv reader (usually into a dataframe type object) which makes it very easy to work with. Databases in my experience are laborious to set up and I would need to learn the API rather than just println("a,b,c,d").