🎣 From a complex graph to a hooked audience

---

You're reading Dataviz Universe, a weekly newsletter about data visualization by Yan Holtz. See all the previous issues here!

---

👋👋

Last week I was working with a client preparing an important presentation. They had a bunch of interesting but complex data and needed to tell a great story with it, so they asked me for advice.

I mentioned a tiny technique I love using. It really resonated with them, so I thought I'd share it here too.
​
​Note: it resonated on LinkedIn too!

😵 The overwhelm problem

There's a very common issue when people build talks around data: the graph they show is complicated and needs a lot of explanation and time to process.

Even if you use all the right dataviz techniques (smart color highlighting, a powerful title, clean annotations..), your graph will often still be hard to swallow.

Take this one. It's a bit old. I showed it at a conference back when I was working on mental disorders and comorbidity. It maps how the major types of health disorders tend to co-occur.

​
When the audience sees a graph like this, there are 2 possible outcomes:

Curiosity: they try to figure out what's going on, stop listening to you, and miss your story.

Discouragement: they give up because it's too complex, and look at their phone instead.

Neither is what you want. 😔

🪄 The fix

What I love doing is revealing the graph step by step.

Start with the setup. Use the beginning to explain the material and method. "This is the X axis: we measured over the last 90 days. This is the Y axis: this metric, measured with this method."

Then show just a few data points. One group only. Tell the story about just them.

Then reveal all the data points.

And finally, maybe reveal a new way to color them that surfaces an insight.

Instead of overwhelming the audience, you build suspense and excitement in the room. People actually follow your story.
​
In my talk, I revealed it piece by piece.
​
First, how the diseases are organized (that's the outer circle). Then how some are far more common than others (that's the bubble size). Then a single connection, to show the strongest co-occurrence between two diseases. From there I relaxed the threshold step by step, adding more links each time, until the full figure appeared.
​
​

That's data storytelling, instead of a list of bullet points sitting next to one overwhelming graph!
​

​
💚 A sentimental example

I also used this technique for my master's graduation. (I know, that was a long time ago 🙈).

I was working on 3 varieties of apple trees, measuring 3 main variables. So I built an animated, rotating cube, explained the variables one by one, then made each population appear while I told its story.

It probably took me longer than writing the actual dissertation. 😅

But the jury loved it, and I walked away with one of the best grades :)

🚀 Give it a go!

This technique is very easy to implement. Sometimes I just copy paste my slide with the graph, then drop white squares on top of parts of it and remove them one by one to reveal it step by step.

Use it with parsimony though. A simple line chart with 1 line doesn't need a step by step reveal.

It's your job to find the right pace. Too slow and the audience falls asleep. Too fast and people lose the thread.
​
What do you think? Inspired to try this next time?
I would love to know!
Please hit reply, I read and answer to everyone!

Have a wonderful day,
​
Yan
​
PS: the Pacific Dataviz Challenge is on (and I'm in the Jury). You should participate!
​
PPS: a spring promo is coming for my Python Dataviz course!
​
​
​