JupyterCon 2023

For the last 3 years, my team and I have been working with some of the largest companies in the world to help transition spreadsheets users to Python. We consistently see data science teams launching JupyterHubs for their business users, but struggling to get adoption from users who rely on spreadsheets.

I will talk about best practices in how to onboard spreadsheet users to Jupyter. This will include how to distinguish the mental model of a spreadsheet and a Python notebook as well as how to select the best spreadsheet workflows to transition to Python.

Then I will go into the reasons that data science leaders want more business users on their JupyterHubs. These primarily fall into three categories:

1. Jupyter allows business users to work with much larger datasets than they can in spreadsheets
2. Jupyter allows businesses users to process data much faster than they can in a spreadsheet
3. Jupyter provides traceability and auditability for your analysis that does not exist in spreadsheets

Next I will talk about all the great open source tools available in Jupyter that can be used to help someone make this transition. These will include:

1. Openpyxl/Xlwings– a python library to read in and/or write Excel files with Python
2. Mito (I am one of the creators) – a spreadsheet interface for Python, inside Jupyter. Each edit in Mito generates the equivalent Python.
3. Lux – automatically create visualizations from your DataFrames, without needing to write Python
4. Streamlit – easily create data apps on top of your notebook

These packages cover the main categories of Python use that data science leaders want to see from their business users: ingesting data, transforming data, visualizing data, and creating apps on top of data.