Manuals and Guides and Programs

We're going to be using the Enthought distribution of Python in this class for Windows and MacOS X computers. Click on the "Subscribe and Download Link," then "Academic Versions" (at bottom of the page). Provide your email, and you'll get your software.

Instructions from Summer 2006

(Most of this is probably out-of-date, but is left here for examples.)

We will do most of our plotting using MatPlotLib. It produces high-quality graphics and is designed to work much like Matlab. Basic tutorial: matplotlib-tutorial.pdf Detailed users guide: matplotlib-users_guide_0.1.pdf

We will be using SciPy (Scientific Python) for much of our processing in the class. It's a collection of enhancements for Python to support scientific and engineering computing. One of those enhancements is NumPy – Numeric Python. NumPy is designed to be highly compatible with Matlab. NumPy-for-MATLAB.pdf. Here's the general manual for NumPy numpy.pdf.

OPTIONAL: The Enthought distribution of Python includes a plotting system called Chaco whose strength is creating interactive graphs, e.g., click on the state in the left window, and see a graph of data for that state on the right. Matplotlib will work just fine for us, but if you want to explore, you're welcome to use Chaco in this class. Chaco_Users_Guide.pdf. Powerpoint slides with a Chaco example: scipy05.ppt

Programs

csvfile.py
fancierplot.py
us_uk_pop_plot.py
bpAvg1990.py
bpStdDev1990.py
counttext.py
viztext.py
countExpression.py
correl.py – This is the code that we did live in class on Thursday 13 July.
us-pop-djia.py – Considering if the US population correlates with the Dow Jones Average
us-pop-djia-nextyear.py – Is it off by one year?
regress.py – Multiple regression in Python
- The first version that I uploaded was broken. The new one will download as regress.1.py, but rename it as regress.py before you get started.
linregress.py – Simple linear regression in Python
testing-multiple-regression.py
Inferential Statistics in Python.ppt

Macintosh Support

Download and install ActivePython (you'll have to register, but it's free). It's an updated version of Python with good things loaded in already (not as much as Enthought, but still a lot).

ActivePython gets installed in a somewhat odd place. To run it, from the Terminal, type /usr/local/bin/python

Then, download and install the ScipySuperpack-ppc-10.4.dmg (or the Intel version, if you have a new MacBook) from http://homepage.mac.com/fonnesbeck/mac/index.html. This includes everything you need for MatPlotLib. You must be running from OS X 10.4 for this to work!

The ScipySuperpack includes four different mpkg's. Double click and install each of them in order: NumPy, Matplotlib, SciPy, and PyMC.

If you don't have Freetype on your Mac, you'll need to install it. (I had it already, for use with LaTeX, so I didn't realize it was needed. Sorry!).
- Go to http://ii2.sourceforge.net/ and download the i-installer. Install it and start it.
- Choose to see the packages available.

Install freetype-2.

Then start up Python (from /usr/local/bin/python), and type from pylab import * and hit return. Ignore the errors you get. Then type plot([1,2,3,4]) and hit return. Again, ignore the errors. Then type show(). Poof! There's your plot! (Note that you'll have to close the figure window to get Python to respond again.) Control-D will end your Python session.

I found that restarting the computer led to all the errors going away.