ITEC696: Python Programming for Business Analytics

Welcome to the course home page!

The steps below describe how to install Python on your machine, including an IDE (Spyder) and the course repo.

Viewing Options

• Option 1: If you are viewing this on Github, you should be able to start viewing the book by clicking on contents.ipynb.
• Option 2: Once you download the course repo, you can read these notes as a browser-based book by opening contents.html. In that event you can navigate forward and back using your browser.
• Option 3: If you also have Jupyter Notebook, you can read the .ipynb files and try running their content.

Installing Python and Spyder

Install Miniconda (preferred) or Anaconda. Anaconda comes with many packages already installed, while Miniconda is a more lightweight version allowing you to install only the packages you want. Here we prefer Miniconda since we'll soon install our own packages.

To install packages we'll use the Python package manager pip (short for "pip installs packages"), which runs on the command line. We will use both Anaconda and pip via the command prompt. While the integrated development environment (IDE) Spyder offers a way to install packages, using the command line will allow us more flexibility to manage virtual environments, self-contained versions of Python. It also practices skills used in remote computing, i.e. doing computations on machines that are far away (often more powerful machines like desktops or supercomputing clusters).

Install Spyder by opening a command prompt (Windows: Anaconda command prompt; Mac/Linux: Terminal) and running

conda install -c anaconda spyder

Spyder is an Integrated Development Environment (IDE) named so because it brings together several software tools that are useful when programming in Python. The two main tools are

• Text Editor (default left pane)
• Python / IPython Console (default bottom right pane)

There are other helpful tools, too, like a Code Profiler, a Help Viewer, and a Plot Viewer. You can find them under View -> Panes.

Virtual environments

One of the key benefits to Anaconda is that it manages virtual environments for us. A virtual environment is a self-contained copy of Python with its own unique packages installed. Virtual environments are useful for managing multiple projects and making it much easier to reproduce code across different machines.

To make a new virtual environment using Anaconda, open a terminal and run the following:

conda create -n itec696 python=3.6

This will make a new directory called itec696 (or whatever name you use) and copy the base Python 3.6 installation within Anaconda there.

---

Now we'll run a quick experiment to see where the virtual environment lives and test whether it works. Try running the lines below in the terminal.

source activate itec696

You should see your command prompt change

    user@machine-name$ |

has now become

    (itec696) user@machine-name$ |

The extra (itec696) signifies the virtual environment is active and commands like python and pip refer to the copy located in the virtual environment.

If you have Windows and your command prompt is slightly different, don't worry. We'll standardize our terminal next. This will give us practice working with a Unix-style command prompt, common to virtually all computers.

Download Cygwin Terminal (Windows users)

Run the setup file and choose a mirror. When you get to the package selection prompt, use the search bar at the top to type in the following package names. For each package, search for the matching name, description, and category, then click the package to select it.

• git: Distributed version control system Category: Devel. For file version control.

• openssh: The OpenSSH server and client programs. Category: Net. For remote computing. (Optional)

• vim: Vi IMproved Text Editor. Category: Editors (Optional)

Terminal Walkthrough

We'll work in the terminal (aka command line) frequently through this course. While many new programmers find the terminal daunting, it's extremely useful, and the simple operations will be familiar to anyone who has used a graphical file explorer.

Like the file explorer, the terminal has a current directory where it's operating. From the current directory, one can navigate to other directories, make new folders, copy or move files, delete items, and so on.

To start, try typing these lines, which use some of the common commands. A description appears next to each command as a comment, a segment of code that is ignored. In both terminal and in Python, comments begin with a hashtag # and can be an entire line or a partial line.

pwd # show the Present Working Directory (PWD)
ls # LiSt the contents of pwd
echo "This is a test" # print text
cd ~/Code # Change Directory to ~/Code
mkdir test # MaKe DIRectory
cd test
ls
cd .. # Change Directory to go up one level
rmdir test # ReMove DIRectory

Some commands above overlap with the usual Windows command prompt, while others have synonymous functions (e.g. ls on Unix is dir on Windows).

Where do these commands live? Use the which command to probe each one. Underneath your input, the terminal will print the directory (i.e. folder) where the command lives.

which ls
which echo
which cd
which rm
which cp
which mv
which mkdir
which rmdir
which git # should work if git is installed
which ssh # optional, if you installed openssh
which vim # optional, if you installed vim

What do the paths have in common?

Finally, let's make a couple symbolic links (i.e. shortcuts) with command ln. These will save us from typing out long Windows paths like /cygdrive/c/Users/$USER/ all the time.

ln -s /cygdrive/c/Users/$USER/Code ~/Code

You can consider adding other folders like Documents, too.

Installing libraries#

In addition to python, Anaconda also comes with a file called pip. Pip - short for "Pip installs packages" - is a package manager that handles downloading and setting up packages hosted on the Python Package Index (PyPi).

Pip can handle installing packages one-by-one (e.g. pip install numpy) or from a text list (e.g. pip install -r requirements.txt). Typically we'll use the latter option when setting up a virtual environment. To install the requirements included in the course repo, try running the sequence below.

pip install -r ~/Code/ITEC696/requirements.txt

You should begin to see lots of output printing to the terminal as Pip downloads each package then installs it. This may take several minutes to finish, depending on your internet connection and processor speed.

The command pip freeze will print the list of currently installed packages. So, if you happen to install a new package and want to generate a revised list of requirements, you can export the output of pip freeze to a file like so:

pip freeze > requirements_v2.txt

Testing Python#

Let's now ensure our libraries load correctly. You can run the test script in Python from the command line:

python ~/Code/ITEC696/test.py

It should generate some output:

Starting imports...
Done!
< DATE AND TIME >

Setting PYTHONPATH (Optional)

If you skip this step altogether, you can still write Python code. You'll just need to comment out any of the Table of Contents cells when running notebook.ipynb files.

We can always add more folders to the PYTHONPATH later by appending them to the variable we just assigned, so long as we separate the folders with a colon :, similar to the system PATH variable.

---

Windows

• Option 1: Open Spyder. From the user menu, navigate Tools -> Current user environment variables. Add key PYTHONPATH with value C:\...\Code\ITEC696\.

• Option 2: Press the Windows key and search "environment variables". Click Set user environment variables. Then add key PYTHONPATH with value C:\...\Code\ITEC696\.

Additionally, we'll need to add this folder to PYTHONPATH through the Spyder GUI under Tools.

---

Mac/Linux

On Unix systems you can run the following, substituting ... for the written out path on your system. Note the usual home folder shortcut ~ may not work here.

echo 'export PATH="/.../Code/ITEC696/unit5:$PATH"' >> ~/.bashrc
echo 'export PYTHONPATH="/.../Code/ITEC696"' >> ~/.bashrc
source ~/.bashrc

Testing Jupyter Notebook

Now that our Python environment setup is complete, we will test one other tool we'll use in this course: Jupyter Notebook. Jupyter Notebook is an interactive, browser-based Python interpreter and is extremely useful for exporting documents.

In your terminal run

source activate itec696 # may be optional, depending on system
cd ~/Code/ITEC696 
jupyter notebook

And watch as a browser window opens, beginning in the same folder where you launched. You can now both

• Start a new notebook Untitled.ipynb by clicking New -> Python 3 from the dropdown menu in the upper right corner.
• Open any of the course notebooks, e.g. example.ipynb or unit0/lab.ipynb.