MacResearch

I was introduced to Python in a three-day tutorial class this summer. Since then, I have found that Python has greatly reduced the amount of work I've had to do, and allows me to automate a lot more tasks--which will be helpful not only for me, but others who will be able to use the scripts in the future.

I think there is one important advantage available from using matplotlib, and other text-based approaches, over graphical approaches: traceability. You don't have to worry about being able to reproduce all the features of the graphs the next time you need it--you just run the Python script or commands again, and you get the same graph.

I also like the extended control over the plotting environment. The ability to specify subregions so precisely, and control where axes do and do not appear, and so on, is quite helpful, particularly if you want to go beyond the traditional single "x versus y" plot.

If you're going to do 3D plotting or things like that, other packages may be more useful; but they're also available.

I use happily matplotlib for all my 2D plotting.

For 3D I suggest mplot3D (mplot3d tutorial). It is included in matplotlib and allows simple 3D plotting.

For more complex 3D needs, I suggest mayavi2 that uses VTK. A simple scripting interface is available mlab.

All of this is available through fink.

I'm using Python more and more these days as I'm migrating away from MATLAB for new work. However I find that GNUPlot is very flexible for plotting compared to some of the python-based plotting solutions I've tried, so I get Python to write out the GNUPlot files for the result sets I need.