Project Jupyter for Productive Data Science Collaboration

Description

Project Jupyter contains tools that are perfect for many data science tasks, including rapid iteration for data munging, visualizing, and creating a beautiful presentation of results. However, the same tools that give power to individual data scientists can prove challenging to integrate in a team setting. Challenges stem from the need to peer review code, to perform quality assurance on the analysis itself, to allow for easy reproducibility and collaboration, and to share the results with management or a client expecting a formal document.

This talk will include sections expanding in detail on the following tools:

• JupyterHub
• Jupyter Notebooks (including templates and custom styling)
• Jupyter Lab
• nbdime
• conda/anaconda.org
• git/github

As data science consultants, we generally work in small teams on our client’s hardware. We have tested and run many different scenarios to work through the best workflow for data science teams and are constantly tweaking the process. This talk will expound on our current setup, the challenges we still face, and how we work around roadblocks.