Skip to main content

Python

You shouldn't need to install python - both your base system, and your conda environment should have a version of python installed already. Check by typing

python --version

Note. On an M1 Mac, the python executable doesn't exist - you have to type python3 instead.

On my system this says: Python 3.9.6 if conda environment is not activated, and Pythong 3.9.13 if it is. This illustrates an important point: if you've installed conda you will have multiple python installations on your system. Also, note that python comes in two similar but slightly different flavours (python 2 and python 3). For the most part you should use python 3.

## Running python

You can run python like this:

python

which takes you into the interactive python command prompt. It looks something like this:

Python 3.9.13 | packaged by conda-forge | (main, May 27 2022, 17:00:33) 
[Clang 13.0.1 ] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>>

Let's try it out, for example we could create a string variable:

a = "Hello, this is a string"

...split into words....

words = a.split( ' ' )

...and compute their lengths:

list( map( len, words )

This little example already illustrates a bunch of stuff in python:

  • split is a string "method" (that is a function, but one which called in the form <object>.<method()>)
  • split produces a list - an ordered list of objects.
  • We also used a so-called functional style: we applied map() to apply a function (len()) to each element of our list.

ipython​

The python interpreter is ok, and it's good for running python programs, but for interactive data analysis work you will probably want to try out something more full-featured: ipython. Quit python (e.g. by typing quit() or pressing ctrl-d) and install it:

mamba install ipython

And run:

ipython

It looks pretty much like before, but you'll have a prompt that says In [1] instead of >>>. It has helpful features like tab-completion of commands and filenames and syntax colouration.

python for data sciecne​

To get a relaly good data science experience in python, we need a few packages. The key ones are numpy (which provides good support for matrices, arrays, and numeric manipulation), pandas (which provides data frames), and matplotlib which is a plotting library.

To install these, use conda:

mamba install numpy pandas matplotlib

Once installed, start a python session and let's try it out with some examples similar to those we ran in R. First import the needed bits of libraries:

from matplotlib import pyplot
import numpy
import math

These statement import the libraries and set them so they can be used locally as pd, np, plt etc.

Now, just like in R, we can make a simple plot:

x = numpy.arange( 0, 2 * math.pi, 0.01 )
pyplot.plot( x, numpy.sin(x) )
pyplot.xlabel( "x" )
pyplot.ylabel( "sin(x)" )
pyplot.show()

img

Or generate and histogram a million samples from a Gaussian distribution:

pyplot.hist(
    numpy.random.randn(1000000),
    bins = 100
)
pyplot.xlabel( "x" )
pyplot.ylabel( "count" )
pyplot.show()

img

Next steps​

To really make interactive python analyses fun, you're probably going to want to install JupyterLab. See how to install Jupyterlab.