A notebook for Excel and Google Sheets
Computational notebooks have taken the DataScience world by storm over the last few years - and for a good reason. They enable transparent and reproducible sharing of findings in data, calculations and analysis through a guided narrative and flexible presentation. They are ideal for Data Scientists to exchange their findings, elaborate on each others’ work and collaborate.
They are also increasingly being used to present findings to a wider audience, including the scary and unpredictable“business side”. However, as Dan Lester pointed out in an insightful Medium post last year, notebooks — the way they are traditionally presented — are not a great way to deliver insights to people who aren’t comfortable with code.
It’s not that business people aren’t data literate. These are people that know their spreadsheets and are often capable of quickly building complex and insightful models there. And it’s not like business users don’t understand analysis and “coding”. As we’ve previously covered — spreadsheets are programs. Excel is a highly productive environment for certain type of development, and there are more spreadsheet developers on the planet than developers of any other language — probably by a factor of 100.
It’s the nerdy looking code blocks that throw them off. And even if monospaced text alone doesn’t send people running off in fear, without knowledge and understanding of the language they cannot read, internalize and thereby gain trust in the underlying analysis.
They are used to being able to “touch” the data (the number of times I’ve heard spreadsheet users use the word “tangible” in relation to data is remarkable), dive into the formulas and play around with assumptions and inputs themselves.
So, while I totally agree with Dan that being able to hide the code cells improves the legibility and decreases the “scare” factor of current notebook environments, the bigger issue is that the audience is unable to dive in and understand how the analyst arrived at the results. Not that they always would, but they want to know they could.
It is definitely good that more and more business people are learning languages such as R and Python. It is also true that for some of the more technical young professionals Jupyter is increasingly becoming “the new Excel”. But Excel is still the main Excel — and will be for a looong time to come.
What the “business side” needs are Notebooks for Excel (and Google Sheets). A web based“presentation layer” on top of the spreadsheets people already have and knowhow to build. An environment that brings the benefits of the guided narrative and presentation of notebooks to their world — not only as consumers, but also as authors. Such a notebook environment will facilitate the same level of trusted and effective data communication and collaboration between business users as notebooks in Jupyter, RStudio, Databricks and Observable do for Data Scientists.
Such an environment must differ from current notebooks in two major ways:
- It has to offer a less technical user experience than current notebooks, both for authors AND consumers.
- It must cater to audiences and groups of collaborators that are a couple of orders of magnitude bigger, as they break away from the realms of relatively small Data Science teams into the ranks and files of every-day knowledge workers.
- I’ve been a bit shy — perhaps too shy — to use this positioning for GRID. And I still doubt that we’ll ever lead with it, but this is in fact what we’re building: a notebook for spreadsheets.
And the success of notebooks in the world of DataScience is a part of the reason I’m so excited about the opportunity for GRID in the world of business.
This article has also been published on Medium.