Learning Jupyter with Bloomberg

This webinar will run Python code in JupyterLab. Therefore, as a basic requirement, you will need to have both Python and JupyterLab installed on your computer. Fortunately this is made easier thanks to package managers like Conda and Pip.

Local or online

You have two options when it comes to installing and running a working Python environment:

1. Use a custom, local installation
2. Use an online, pre-configured environment

We recommend following along with a local installation, as it achieves two main goals: it teaches an integral part of development and Python programming in general, and it gives you more control over the installed packages and your data (which will persist in a local installation, but might not in an online environment). The instructions below will cover both options.

Conda and Pip

Conda is a language-agnostic cross-platform environment manager. Pip is a general-purpose manager for Python packages.

For you, the user, the most salient distinction is probably this:

• Conda installs any package within conda environments;
• Pip installs Python packages within any environment.

We highly recommend using Conda as your package manager for this webinar, as it will allow you to install non-Python dependencies in an isolated environment. This will make for a much smoother installation process! But, if you still want to use pip, we have provided installation instructions below.

For Conda installation: https://docs.conda.io/projects/conda/en/latest/user-guide/install/ For Pip installation: https://packaging.python.org/tutorials/installing-packages/

Local installation with Conda

# Clone this repository
git clone https://github.com/ibdafna/learning_jupyter_with_bloomberg

# Navigate into the root directory of the clone repository
cd learning_jupyter_with_bloomberg

# Execute the Conda installation recipe file. This will create Conda environment and install all dependencies.
conda env create -f binder/environment.yml

# Activate the Conda environment
conda activate jupyter-masterclass

# Create a kernel for this environment
ipython kernel install --name jupyter-masterclass --display-name jupyter-masterclass --sys-prefix

# Build JupyterLab assets
jupyter labextension install @jupyter-widgets/jupyterlab-manager bqplot

Once you are done with the installation steps, you can start JupyterLab by executing jupyter lab in the root repo directory.

Local installation with Pip

Pip is a Python package manager. As such, it can only bundle Python-related packages such as pandas, numpy, etc. Although Python is a big and integral part of the Jupyter ecosystem, we rely on other languages, such as JavaScript and TypeScript, for all of the front-end code. This means that we work with nodejs and npm quite a lot.

To build JupyterLab and its assets, we need to transpile TypeScript code into JavaScript code and then bundle it into a single file which is consumed by the web browser. This process is achieved with npm, a JavaScript package manager. npm is bundled with nodejs, so to get it installed, you will need to follow the 'nodejs' installation process. Please make sure npm is installed on your machine before continuing with the installation instructions below.

nodejs/npm installation: https://nodejs.org/en/

Once you have installed npm, follow the installation instructions below:

# Clone this repository
git clone https://github.com/ibdafna/learning_jupyter_with_bloomberg

# Navigate into the root directory of the clone repository
cd learning_jupyter_with_bloomberg

# Create a virtual python virtual environment
virtualenv jupyter-masterclass

# Activate the virtual environment
source jupyter-masterclass/bin/activate

# Install all packages
python -m pip install -r requirements.txt

# Create a kernel for this environment
ipython kernel install --name jupyter-masterclass --display-name jupyter-masterclass --sys-prefix

# Build JupyterLab assets
jupyter labextension install @jupyter-widgets/jupyterlab-manager bqplot

Once you are done with the installation steps, you can start JupyterLab by executing jupyter lab in the root repo directory.

Use an online, pre-configured environment

Binder is great for experimenting, but note that any changes you make to these files will not persist, and the environment will time out after a period of inactivity

Click the image below to run Binder with a pre-configured Jupyter Masterclass environment: