The more I advance into analytics, the more I come back to Excel as a teaching and prototyping tool. Yes, of course, Excel has its weaknesses — but as a medium for learning, it’s unmatched.
Here’s why:
It reduces cognitive overhead
Cognitive overhead is described as “how many logical connections or jumps your brain has to make in order to understand or contextualize the thing you’re looking at.”
Often an analytics learning journey looks like this:
- Learn a brand-new statistical technique.
- Learn how to implement the brand-new technique using brand-new coding techniques
- Progress to more advanced statistical and coding techniques, without ever having felt really comfortable with the basics.
It’s hard enough to learn the statistical foundations of analytics. To learn this while also learning how to code invites sky-high cognitive overhead.
Now, I do believe there is great virtue to practice analytics via coding. But it’s better to isolate these skill sets while mastering them.
Excel provides the opportunity to learn statistical techniques without the need to learn a new programming language at the same time. This greatly reduces cognitive overhead.
It’s a visual calculator
The first mass-market offering of a spreadsheet was called VisiCalc — literally, a visual calculator. I think of this often as one of the spreadsheet’s biggest selling points.
Especially to beginners, programming languages can resemble a “black box” — type the magic words, hit “play” and presto, the results. Chances are the program got it right, but it can be hard for a newbie to pop open the hood and see why.
By contrast, Excel lets you watch an analysis take shape each step of the way. It lets you calculate and re-calculate, visually.
Seeing is believing, right?
You can’t take shortcuts
Open-source tools like R and Python give you access to a wide variety of packages, which usually means you don’t have to “start from scratch” with basic functions.
While Excel add-ins for analytics are available, many of them cost. But that’s OK! In fact, left with the bare building-blocks of Excel, there’s more opportunity to get face-to-face with what’s being built.
In Excel, we can’t always rely on an external package to conduct our analysis for us. We’ve got to get there by our own devices.
It forces you to be agile
A temptation in data analysis is to build the most complicated possible model at first, and then work backward to find something that works. It’s better to go in reverse: start with a minimum viable product, and iterate from there.
It’s a lot harder to build a complicated model in Excel than in Python — which is a limitation, when we need complicated models — but as a prototyping tool, this is great, because it forces us to start small.
We’re not making production models here
I just highlighted some of the many benefits of learning analytics in Excel. Can you think of others? Or maybe you’re not convinced?
One of the biggest objections to doing analytics in Excel is that it can be error-prone and hard to reproduce.
That is absolutely true, but we’re just learning here. We’re not making production models.
Don’t discard Excel’s aptitude as a teaching tool for its flaws as a fast, reproducible analytics workflow.
Learning analytics in Excel: What next?
I’ve learned more about statistics and analytics by experimenting in Excel than any other tool, and I hope this approach can work for you too.
If you’d like to see what Excel can do for your learning path, check out my book Advancing into Analytics: From Excel to Python and R. The first part of the book demonstrates crucial analytics concepts and techniques in Excel, then builds on this knowledge in R and Python.
Learn more about Advancing into Analytics, including how to read for FREE, here.
How do you prefer to learn statistics and analytics? Do you see other pros or cons of using Excel? Let’s talk in the comments.
Leave a Reply