Last week, I applied the optimist vs. pessimist style to a map during Office Hours.

(The real version of the map was about a different topic in a different state, but you get the idea.)

The Traditional Map

I’ve written about traditional and storytelling approaches before. You’re probably familiar with this terminology already?

The traditional version would have a topical title and regular ol’ brand colors.

The Storytelling Map: Optimist Version

In this optimist version, we’ve got a:

• “good” takeaway title;
• “good” brand color;
• “good” icon; and a
• call-out box highlighting a superstar.

The Storytelling Map: Pessimist Version

In this pessimist version, we’ve got a:

• “bad” takeaway title;
• “bad” brand color;
• “bad” icon; and a
• call-out box highlighting a laggard.

Which Version Should You Use?

That obviously depends on your audience and context.

You’re welcome to comment here with your own insights. I read every response.

Download These Maps

Want the Excel file that I used to create these maps? You can download it here.

…and you’ve seen heat tables…

…but have you tried tables with bubbles?!

There are obviously pros and cons of each approach – like everything with data visualization.

Comment below with your own insights into the pros and cons.

How to Make Tables with Bubbles in Excel

You’ll set up a Helper Table with x, y, and z values.

Then, you’ll insert a bubble chart.

Finally, you’ll adjust the axes, grid lines, colors, and labels, just like you normally would.

Let me know if you’d like a more detailed tutorial.

Download the Excel File

Or, you can download my Excel file and explore it yourself.

This is just the file I used for this blog post (not step-by-step instructions).

…why not try bubble charts?

Bar charts depict data with length.

Bubble charts depict data with area.

If you want to try bubble charts, you can make them in good ol’ Excel with this tutorial.

The stacked chart wasn’t terrible.

She already formatted it, following all the usual dataviz and accessibility principles. (White outlines, brand colors, direct labels instead of legends, sequential colors for a sequential variable, only labeling the min and max on the scale, using sufficient color contrast, etc.)

The challenge was that her chart was going into a dynamic dashboard, so depending on which filters were selected, the numbers might be really big, really small, or even zero.

We needed to think forward and plan for all those different scenarios.

Challenge: A Single Stack is Lonely

Another challenge was that her chart looked lonely.

I don’t love stacked charts for a single stack.

I don’t have peer-reviewed articles to back this up; it’s a personal preference.

Stacks are extroverts. They thrive with their friends, like this:

Challenge: Stacked Charts are Tough for Tiny Slices

Remember how her chart was going into a dynamic dashboard?

So depending on which filter(s) were selected, the numbers might be really small.

Stacked charts are terrible for small slices. They’re hard to label, and even harder to see.

You’ve run into this, haven’t you? It’s a common dataviz headache.

Challenge: Stacked Charts are Tough for Zeros

Another common dataviz headache is that stacked charts are terrible for zeros.

Because they literally can’t show zeros. There’s nothing to visualize.

Sure, I’ve seen people try all sorts of workarounds, like adding text boxes off to the side that explain that a category isn’t shown because the number was a zero. But that’s messy and time-consuming.

Solution: Use Columns/Bars Instead of Stacks

I often recommend changing the stacked bar/column into a regular bar/column, like this.

(Should you use vertical columns or horizontal bars? I’m using vertical columns here because this particular graph is sequential.)

(Worried you’ll have to delete the chart and start from scratch? No way! It’s faster than that. You can change the chart type with just a few clicks.)

Can you see how regular bar/column charts work better for tiny numbers?

Can you see how regular bar/column charts work better for zeros?

And don’t worry — regular bar/column charts still work when comparing multiple categories, too. We’d simply make a small multiples layout, like this.

Your Turn

These principles can and should be applied to every software program.

As you can see, I’m using everyday software (Excel) in these examples.

Have any how-to Excel questions? Comment here and let me know. I’ll post a link to an existing tutorial, or I’ll make a brand new tutorial for you.

And other times, lower numbers are better.

Here’s how you can add icons and text boxes to graphs to remove that guesswork for your colleagues.

This example was made in Excel – but the techniques can (and should!) be applied to visualizations in all software programs.