Installing R and rpy2 from scratch

Python has seen a bit of an explosion in packages for statistics, machine learning, and data structures to support the two. Between pandas, statsmodels, scikit-learn, and scipy.stats Python developers now have a miniature data stack for econometrics, machine learning algorithms, or statistics, as well as sensible data structures with tons of convenience functions.

Still, those packages may not provide everything you need to get the job done. If, like me, you're a Python dev analyzing some time series data, you might need to visit an old friend with excellent, full-featured statistical packages like caret and forecast.

You'll be happy to know that, with rpy2, you can have your cake and eat it too. Installing rpy2 can be painful, however, especially if you're working with EC2 instances on Amazon Web Services. New instances contain little or no GNU tools, making an R + rpy2 installation harder than

sudo apt-get install r-base r-base-dev python-dev python-setuptools
sudo easy_install pip
for pkg in numpy scipy rpy2; do
    sudo pip install pkg

I've attached a script below to help devs get past this problem. It has been tested to work on Ubuntu 11.10, and I don't expect any compatibility issues with newer versions.

Here's a Gist with the script. If you run into issues, please let me know and I'll try to fix them and update the script.