Python II: Data Analysis and Visualization

Expectations

• Previous Python experience required. Focus is on using Python as a tool, not on learning the language.

• Python is not the solution to all your problems. (I.e., this seminar is informational rather than evangelical in nature. Draw your own conclusions.)

• Python is not a magical gift box; it is a programming language. At least some work will be required to reach your goals.

• Consider me to be a tour guide rather than an expert.

  • Will highlight capabilities of various packages, but have very little experience with most of them.

Useful Resources

• Stack Overflow: http://stackoverflow.com (Someone has probably already asked your question and received an answer or a dozen answers. If not, you can ask questions here too.)
• Python 2 Reference: http://docs.python.org/2
• Python 3 Reference: http://docs.python.org/3
• IPython Notebook Gallery: http://nbviewer.ipython.org/
• IPython Reference: a collection of IPython tutorials

IPython

• Copy and paste of example output.

• Pop-up help.

• Tab completions.

• Automatic pretty-printing.

• Persistent history.

• Logging sessions.

• Demo mode.

• Loading Python files.

• Pylab

  • vs SAGE

NumPy

• NumPy arrays vs Python lists.

• Creation of arrays.

  • array
  • arange
  • linspace
  • zeros, ones

• Reshaping arrays.

• eye

• Element-wise operations.

• Simple linear algebra.

• Simple stats.

Note

If you were in the morning session, then you want to np.resize to resize a NumPy array and not np.reshape as I accidentally wrote. Sorry for the problem.

SciPy

matplotlib

pandas

StatsModels

NetworkX

NLTK

scikits

• scikit-learn
• scikit-image

SymPy

• Running isympy.

• Using SymPy from within an IPython GUI.

  • Example.

StarCluster

Miscellany

• mpi4py
• IPython parallelism
• Cython
• Numba
• PyCUDA