Thread by @rappa753

BEFORE
- Standard bar chart
- No clue what the message is

AFTER
- Bar chart with nuanced color use
- Informs your reader about key insights & actions

The latter is actually pretty easy to pull off.

At the end of this step-by-step guide, you can do that too. #rstats #dataviz

Here's our starting point.

All code is available at albert-rapp.de/posts/ggplot2-…

Note that this tutorial is a ggplot2 recreation of

// Labels on y-axis

First, move the names to the y-axis.

This is important when the labels are real names instead of IDs.

No one likes to tilt their head for reading.

// Start with grey

Use the group aesthetic instead of fill.

Reserve colors for storytelling.

// Reference points

Add a reference point to put your data into perspective.

A simple average can be enough. Here, you see the average as the new category "ALL"

// Order top to bottom

Notice that I also ordered the bars from top to bottom? fct_reorder() did that for me.

This way, your reference point can do its job perfectly.

(Ordering is also a good idea even without reference points)

// Put important bars on the outside

I've also arranged bars with factor(). This way, important quantities are on the outsides (left and right).

This makes comparisons easier.

// Highlight with colors

Use colors to highlight your story points.

PRO TIP: Save your colors as variable names.

Usually, you will try out different colors.

Copy and pasting a new color into multiple parts of your code is tedious.

// Avoid clutter

Remove anything that clutters your plot.

This includes useless grid lines.

Do this by setting expand = FALSE in coord_cartesian()

Notice how I've also set clip = 'off'.

This allows drawing outside the grid panel.

We'll need that later.

// Move and format axes

Move x-axis to the top with scale_x_continuous().

This way, your readers can see it early on.

Also, grey out the axes labels with theme(). Otherwise, they can be distracting.

// Align labels

It's a good idea to move labels such that they form a clean line (e.g. with an axis)

Use theme() to move the axis titles to the top left corner.

BONUS: The reader can see the titles even sooner.

// Align manually

Not all things can be set in theme(). For example, the axes label can only use a single hjust value.

But "clear" lines need
0% label ➡️ left-justified
100% label ➡️ right-justified

An annotation outside the panel grid will do the trick.

// Label directly

Use geom_text() to label your highlights.

// Add a legend

We can add a small legend using another annotation outside the panel.

This will be a richtext annotation because we highlight a single word.

// Create a story plot

The bar plot is finished. Time to create the story texts.

Create a text-only plot first. We'll combine it with the bar plot later.

// Combing bar plot and story

Use {patchwork} to put together your two plots.

Add the call-to-action prompts via a new title + subtitle.

Done! That's a wrap. 🥳

Hope you found this guide helpful. You can follow @rappa753 for more content like that.

And don't forget to jump back to the first tweet below to like or retweet the guide. I'd appreciate it 👌🏻

https://twitter.com/rappa753/status/1575878220372594688

Want more helpful resources?

Every week, my newsletter shares insights on
- dataviz,
- Shiny web apps
- stats

Reading time: 3 minutes or less

You can join at
alberts-newsletter.beehiiv.com/subscribe