Data Storytelling 11: Aesthetics & Acceptance
Pretty privilege exists in data science too
1 - Aesthetics
Should you make it pretty?
Yes.
People generally assume the more pretty designs as being easier to use & read than designs which are not pretty. Surprise, surprise, pretty privilege exists in data science too.
In data storytelling, spending time to make your designs aesthetically pleasing to the eye can mean your audience will have more patience for you (which means you can make more fk ups, and get away with it).
Recall what bowtiedbull once said in his efficiency book:
Stop for a minute, and look at the ads & billboards in your city. Stop, and think WHY it looks pretty.
Same concept applies here. You don’t need to re-invent the wheel, someone else has probably had to tackle a similar problem to what you are doing right now. Take a look at several examples that people have attempted, then study the best ones, and try to bring what they did correct to your graphs.
Generally speaking, here are 3 things you’ll want to pay attention to, when trying to make your visual appealing:
Color: Don’t use color without a real reason. You should use color only when you want to highlight, or differentiate something.
Alignment: you should organize elements on your page to create clear vertical & horizontal lines to establish a sense of cohesion.
Use white space: Keep the margins there, please don’t stretch your graph because you feel like you have to fill the space. (Same advice applies to your resume btw. White space is good, trash buzzwords are not.)
Here’s an example below. If the goal of your graph was to compare categories B & C. Which would you be more likely to look at? The graph on the left or the right
This is partly funny because the graph on the right is better look visually. Is easier on the eyes, and actually takes *LESS* time to make. Here’s a quick bullet list on what the graph on the right does better:
More whitespace = easier on the eyes
Numerics listed onto the bars = reader doesn’t need to look to the left to see the value
B & C are color coded = this means A, D, and E basically fall into the background, the the reader’s eye is grabbed directly onto B & C
The only other thing I’d do with the graph on the right is reduce amount of white space between each of the bars, as this will make it even easier on the eyes.
2 - Acceptance
For your graph to be effective, it must be ACCEPTED by it’s intended audience. Now, here’s a problem, what if the new visual you have isn’t accepted by it’s audience?
Let’s say your team has an old dashboard, or an old visual template. Let’s say they have complaints about it. So, you offer to take a crack at improving it, and you used all of the skills on data storytelling in this substack, and voila you basically solved all of the complaints. Do you think your visual would be accepted by the audience?
No!
Fact of the matter is, people are lazy and they hate change. People generally want to avoid making any changes, as they like doing things “the way we’ve always done it”. They’d rather keep making the same complaints about their dashboards over and over again, than let someone show up & fix the problems, but do it in a brand new way.
If you are tasked with handling the internal dashboards, or the analysis tools, this is something you will encounter. So, let’s talk about how to address this problem.
Articulate the benefits: Some people need to be told WHY things will look different going forward, and why this approach is better. Be sure to give full transparency here.
Side by side example: Sometimes people want to see the before & after side by side. You should combine this with the above as well.
Provide multiple options: Instead of giving people a design, and calling it a day. Instead, come up with about 3/4 designs, and ask people for which one they prefer, and ask for any feedback you can get. People tend to value things that they contributed to more, and voila, this will help get them to accept your visuals.
Get a vocal member on your team: Find a guy who is an influential person at the company, and talk to them on a 1 on 1 basis. If you can basically handle all of this guy’s criteria, he will give the go ahead. And, if he’s influential enough, then everyone else will basically fall in line :)
One thing to keep in mind is that if you are going to be doing the above bullet points, you need to make sure that your visuals aren’t the problem. An easy way to confirm this is to find someone whose not invested at all into your project, and show them a few graphs, and ask them a few questions like:
What do you like?
Any questions you have after looking at the graph?
Is the graph hard to read?
At this point, you have all of the theory on what makes good graphs down. Next post will be entirely about looking at good visuals, breaking them apart, and explaining why they are good.
It’s basically putting the past 11 posts on data storytelling to practice.