MOST DATA ANALYSIS INVOLVES A GOOD DEAL OF DATA MANIPULATION AND NUMERICAL COMPUTATION. OF course, we use computers for these tasks, hence we also need appropriate software.
This appendix is intended to give a brief survey of several popular software systems suitable for the kind of data analysis discussed in the rest of the book. I am mostly interested in open source software, although I also mention some of the most important commercial players.
The emphasis here is on programming environments for scientific applications (i.e., libraries or packages intended for general data manipulation and computation) because being able to operate with data easily and conveniently is a fundamental capability for all data analysis efforts. On the other hand, I do not include programs intended exclusively for graphing data: not because visualization is not important (it is), but because the choice of plotting or visualization software is less fundamental.
In many ways, our choice of a data manipulation environment determines what problems we can solve; it certainly determines which problems we consider to be “easy” problems. For data analysis, the hard problem that we should be grappling with is always the data and what it is trying to tell us—the mechanics of handling it should be sufficiently convenient that we don’t even think about them.
Properties I look for in a tool or programming environment include: ...