There are two main versions of Python available: Python 2.7.x and Python 3.x. At the moment, Python 2.7.x is preferable for data analysis and numerical computing as it is more mature. The pandas library requires a few dependencies in order to be setup correctly – NumPy, SciPy, and matplotlib to name a few. There are a myriad number of ways to install pandas – the recommended method is to install one of the third-party distributions that include pandas. Distributions include Anaconda by Continuum, Enthough Canopy, WinPython, and Python(x,y). Installation of the IPython package is highly recommended as it provides a rich, highly interactive environment for data analysis.

Thus, setting up our environment for learning pandas involves installing a suitable version of Python, installing pandas and its dependent modules, and setting up some useful tools such as IPython. To re-emphasize, I strongly advise readers to do themselves a favor and make their task easier by installing a third-party distribution, such as Anaconda or Enthought, so as to get their environment up and running trouble-free in the shortest possible timeframe. In our next chapter, we will start diving into pandas directly as we take a look at its key features.