NOTE: the installation step is expected to be done prior the beginning of the block course.

Please try to proceed with the installation using Option 1 or 2 below. If failing that, the binder (Option 3) is available and does not require installation.

However, it won’t save the results of your computation unless downloaded. The running session is subject to idle timeout.

Each of the following three options should give you a jupyter notebook server.

Once the installation is completed, proceed to Getting started

Option 1: on your own computer

First of all download the notebooks from the Github page by either
  • clicking Code, Download ZIP, or

  • cloning the GitHub repository with opening a terminal (Git Bash on Windows) and typing:


You will need the following packages:

  • cython (install this first)

  • ultranest==3.2.0

  • pystan==

  • cmdstanpy==0.9.76

  • snowline==0.5.2 (requires pypmc, see this issue)

  • jupyter

To install these with conda:

  • Download and install Anaconda or miniconda

  • Open a terminal (Anaconda prompt in Windows). If conda is not active yet, run conda activate.

  • Add conda-forge repository: conda config --add channels conda-forge. You can check with conda config --show channels

  • Install the packages:

    conda install -y cython ultranest==3.2.0 pystan== cmdstanpy==0.9.76 arviz==0.11.2 jupyter
    pip install snowline==0.5.2
  • In the folder with the notebooks, run the jupyter server:

    jupyter notebook
  • Navigate to the working folder and the open the notebook you want to work on.

The Dockerfile below gives an example of the installation process on Linux.

If you have problems, search online first (stackoverflow, how jupyter notebooks work, etc). You can also open a Github issue

Option 2: with Docker

Download the notebooks from the Github repository (see option 1).

  1. Install docker

  2. Download and place this text file (called “Dockerfile”) <> in a empty directory. (You can also download it from the github repository)

  3. Open a terminal and enter this directory.

  4. run:

    docker build --rm -t jupyter/my-datascience-notebook .
  • This will build a docker container (similar to a virtual machine) with all the python packages needed for the course.

  • Its name (tag) is jupyter/my-datascience-notebook.

  1. run:

    docker run -p 8888:8888 -v "${PWD}":/home/jovyan/work jupyter/my-datascience-notebook jupyter lab --ip= --allow-root
  • open the stated URL…. in your browser

  • This will run a JupyterLab server, where you can edit notebooks. Here some explanation of the command above:

    • -p 8888:8888 allows the container to use the web server port 8888 on your computer

    • -v makes the current directory accessible as the work/ directory

    • jupyter lab … starts the jupyter lab server

  1. Work with notebooks in the work folder, you can upload the downloaded notebook to there. You do not need to install anything anymore (remove first two cells).

Option 3: Binder

After installation

Once the installation is completed, proceed to Getting started, to get access to the Zoom information.