I recently do a lot of mathematical computation, using MATLAB, unfortunately MATLAB is not open-source and fairly expensive to buy. There is a open-source reimplementation of the MATLAB language called GNU Octave. But I never really liked it, mostly because of the missing visualization features compared to MATLAB.
So I started to look for something else, and as I’m a huge fan of Python it was obvious to take a look at SciPy, a Scientific extension based on NumPy, a highly optimized multi-dimensional array library. And I must admit I don’t regret it.
Basically my environment consists of several components:
Python - The rock every thing is built upon
NumPy - The fast multidimensional array/matrix library
SciPy - Provides most functions MATLAB does
IPython - An interactive shell with many great features
matplotlib - Visualization of the data in a way similar to MATLAB
To get the most convenient environment I start IPython in the pylab mode by calling
$ ipython -pylab
If you are used to MATLAB and want to give SciPy a try, take look at Mathesaurus, especially at NumPy for MATLAB users. It’s a good way to get quickly started with Python as a replacement for MATLAB.