Install scipy, numpy and pandas OS X Mountain Lion

Since I last posted a “how-to” blog-post regarding the installation of
numpy, scipy, and pandas, things have gotten a good bit easier, so
this is going to be a shorter post.

I’m working with a fresh install of Mountain Lion OS X, with nothing
else installed.

1. step 1: Install homebrew for mac, using the
   ruby script found on its homepage:

    $ ruby -e "$(curl -fsSL https://raw.github.com/mxcl/homebrew/go)"
   

2. If you follow the homebrew instructions at the end of the install,
   you should then run

    $ brew doctor
   

   and get a complaint about not having X Code installed. At that
   point, go the App Store, download X Code. After you download X Code,
   go to Preferences and download the X Code Command Line Tools
   under Components.

   At this point you should be able to run: \$brew doctor

   and get the encouraging message:

    your system is raring to brew!
   

3. Now armed with homebrew, and a quick glance at Homebrew & Python
   Page you
   can see that Python 2.x is recommended to be installed via:

    $ brew install python
   

   if you’re installing 3.x you can see replace the word python
   with python3 and you’re good to go. Make sure things are done this
   way, that way you automatically get the install of distribute and
   pip (which will be used in the next steps)

4. At this point, use brew to install any other packages you’d like
   (for instance, git was a no-brainer along with a couple of
   others).

5. One package that you MUST install with homebrew is gfortran,
   as the scipy library depends on this. Therefore, the next step is:

    $ brew install gfortran
   

6. After that’s done, you simply need to:

    $pip install numpy
    $pip install scipy
    $pip install pandas
    $brew install freetype #see my last post if you want to know why
    $pip install matplotlib
   

7. If you wish to test your installations, you need only

    $pip install nose  #needed for all tests
   
    #inside of IDLE or iPython
    [In:] import numpy
    [In:] numpy.test('full')
    [In:] import scipy
    [In:] scipy.test()
   
    #from command prompt
    $ nostetests pandas
   

And there you have it… happy key tappin’!