Why a Graph is Not Enough

A picture can tell 1000 words but not always the words you intend

Next slide, and TA-DA - here is a graph! Typically considered the pinnacle of any technical presentation, a good graph is always well received. Or is it?

Today, I address one of the most common mistakes STEM professionals make when presenting their work. It is not working their graphs enough.

The fact of the matter is that people make their own sense of information presented in the form of a graph, and they do it fast.

Neuroscientists from MIT postulate that we can process entire images that the eye sees in as little as 13 milliseconds [reference link].

Our brains are to blame. Their primary function is to make sense of the world around us, simplifying complex ideas into digestible bits. Brains love patterns - so much so that they often see patterns where none exist. Brains love to fill in blanks, to connect the dots, and to create a more coherent view of the whole. That’s why presenting graphs without walking your audience through them is risky - it leaves big room for misinterpretation.

Take a look at the very simple graph below. Left unguided, your audience will interpret it in their own way, focusing on different aspects depending on their experiences. Many will focus on the main trend, some will fixate on the outliers, others will question the scale or spread of the data, and a few might not even notice the units. Everyone will make some sense of your graph, but you want them to understand your insights and takeaways, right?

You Have a Graph to Show - Here’s What to Do

Before I get into the step-by-step guide, remember: you worked hard to create this graph, and you alone know it inside-out. Place yourself in your audience’s shoes - they have no clue! Yes, even if they are experts in your field, they won’t know what kind of results you’re about to show or how you chose to present them. To ensure your audience sees the story you intend, walk them through it.

Your Step-by-Step To-Do List

1. First things first: the axes and the units.

• Explain what’s on the X and Y axes.

• State the units clearly. Depending on your field, you may even need to define your units.

2. Outline the scale.

• How far does your graph stretch on the X and Y axes?

• Is your scale typical for your field? Highlight anything about the spread of data that adds context or demonstrates your expertise.

3. Present the data.

• Walk your audience through the data points before introducing a trend line.

4. Show the error bars.

• If you don’t have any, explain why. This is a great moment to share insights about uncertainties and cumulated errors - whether they stem from calculations, statistical methodologies, or empirical results.

5. Add the trend line and interpret the results.

• This is your time to shine! Share your observations and ensure they’re crystal clear. Your insights are the true added value of your work.

• Pro tip: Use short text bubbles, highlights, arrows, or circles on your slides to guide your audience’s eyes to the main takeaways.

6. For multiple data sets:

• Present each set one at a time, repeating steps 3-5 for clarity.

• Add a legend as soon as the second data set appears.

P.S. Make sure your audience can see every element, regardless of where they’re seated. As a rule of thumb: go BIGGER and bolder with fonts (for your axes, units, legend, title, data points, error bars, etc.).

P.P.S. Use animations strategically so elements appear as you speak, leading the audience through your thought process step-by-step.

P.P.P.S. Avoid blurry graphs or irrelevant parts. Use high-resolution images (at least 400% clarity). If you don’t plan to discuss a specific part of the graph, leave it out entirely - your audience deserves clarity, not distraction.

Final Thoughts

Love your work and love your graphs! If you want to see a true master at work, check out Hans Rosling presenting his data pool. Watch how he literally kneels before his graph!

Let me know in the comments what you think is the most important thing when it comes to graphs.