Oh no! I think my project is outgrowing my Jupyter notebook. How do I survive?
Interactive notebooks, whether Jupyter, google colab, or others are fantastic for lots of things. A range of options to visualise results, advanced markdown use for communicating process and interpretation, and simple to learn for newbies. But they also get a bad rap for encouraging poor programming techniques which make notebook-code hard to reuse.
As a recent notebook-trained data science bootcamp grad, I’m excited to share what I’ve learned from my move into the real-world of organisational data about the strengths and challenges of notebooks. I’m not an evangelist for notebook lovers or loathers - just someone figuring out how to use tools to make things that work. Hopefully this talk will help you think through the choice to notebook or not more intentionally.
An example-focused discussion of the pitfalls and strengths of interactive Python notebooks. Topics up for discussion: Namespace pollution - what is it and why do I care? Speed - when might my notebook be holding me back? Pretty pictures - are there times when even for visuals I might skip the notebook? Transitions - I’m so comfortable using my notebook. How do I change in useful ways without grinding to a halt? Functional combinations - can there be a happy middle ground between notebooks and scripts?
Lydia is a data wrangler who brings a diverse background to understanding what numbers tell us about the world. She joined the Data and Analytics team at Neighbourlytics after completing a data science bootcamp last year. Prior to that she provided training and coaching for data-driven change in STEM education, and led rock climbing and kayaking adventures for people of diverse backgrounds and abilities, among other bits and pieces.