Sharing how we can communicate our science effectively, whether in posters, papers or on social media!
Share
SciCom – How To Design Figures
Published 4 months ago • 5 min read
Graphical Design Fundamentals
Hi Reader, do you change the line thickness of your graphs?
I ask because how we present our data has a tremendous influence on how it is perceived.
There's neither objective interpretation nor representation of data.
Therefore, let’s deliver you the fundamentals of effective design:
Design Matters
Remember last week's lesson - we saw how proper design ensures that our readers can accurately process our figures.
Second, it helps them understand our results quickly so that, with the limited time they have, they don’t skip over our work but instead cite it.
Displayed are mean data ± SEM with 6 data points. Obviously, the graph on the very left makes it extremely hard to focus on the data. However, even the center one is visually distracting, as the y-axis label is much too large, and the strong black–white contrast of the bar filling makes it hard to see whether one bar is higher than the other. Although I think we could even optimize the one on the very right by adding data points and switching to SD, it overall allows for the best assessment.
Poorly designed figures can make readers doubt whether your work is solid. This is also why some journals like Nature have guidelines.
Let's therefore see how to allow our readers to grasp all information quickly and without getting distracted.
Axis Dimensions
Once you compiled your data and chosen a figure type to display it, double-check whether your statistics program automatically chooses the range of your y-axis.
Apart from showing whether there is a difference in the mean, the graph on the left subconsciously suggests that there is “room to the upside.” Even though there is no data there, it implies to our mind that there could or maybe should be data points. And maybe there are: if we don’t plot individual values, there could be.
If your y-axis range is too large, differences between groups become visually harder to see, potentially misleading your audience.
Next, think about the dimensions of your graph in terms of the x-axis. Make sure your readers can visually assess all data points quickly.
When it comes to curves, consider that the spacing of your x-axis has an important impact on the perceived slope of your graph. Again, there is no way around this. Also consider that the perception of your error bars will vary with the size of your graph. In both cases, mean values ± SD are displayed.
If you squeeze them too close together, it will become difficult. If you stretch the axis too long, it becomes harder to compare data points and differences.
Axis & Data Labels
Your axis labels should allow readers to immediately understand what is being displayed - even without reading the description or main text.
This is especially important because many readers simply scroll through figures to judge relevance.
It was hard to come up with a y-axis label example that makes sense… anyhow, what I want to demonstrate is that A) excessively big labels seem odd even though you might think it helps readability, B) if your journal hasn’t got a corporate design, choose a proper font (not fitting Arial Black on the left, better Yu Gothic UI Semilight on the right - it doesn’t need to be special, but choose something simple yet modern), C) “Threshold Count” is very unclear, and even if you need two lines, go for something understandable. By the same token, name the treatment or sample type in your legend properly. Sometimes you cannot avoid abbreviations, but perhaps you can choose ones that can be “read,” e.g., “HEK-IL2 KO” (cell line - proten/gene - treatment) instead of shorter yet cryptic labels like “HI2K”.
Pay attention to where you place your legend when you have large panels. Don’t underestimate how long it feels when your eyes have to travel across several charts just to double-check the legend.
However, always include a legend in the panel, not just in the description:
Click to enlarge. Admittedly, this paper from Woodhams et al. is from 2001, but if you have the legend only in the description, it can be very frustrating. This is especially true in this case, where in the original there were eight graphs per page. However, using sensibly spaced horizontal dotted lines really helps to follow up on the data. If readers must search in the description or even in the text to understand what you did, they are more likely to skip your manuscript.
In complex setups, as seen in many Nature papers, it can be helpful to add an additional panel showing the experimental setup.
Axis Ticks
When defining axis ticks, aim for clarity rather than minimalism or maximalism.
Ticks should help readers understand where bars or data points lie.
One important point to note: if you set the y-axis manually, it can happen that you end up with a “blunt” end, as in the first and second panels. Double-check whether this happens and whether it poses a problem in your case. Moreover, as you see in other figures above, although not very common in publications, I strongly recommend considering dotted helper lines - they make it much easier to judge where bars or data points end and whether they exceed certain thresholds.
Importantly, ensure all text elements - tick labels, axis labels, legends - are large enough to read easily.
Line Thickness & Patterns
An often overlooked topic of utmost importance is line thickness. Several scientists might not even know that they can edit it in most software.
Whereas in the upper figure the bold lines are so thick that it becomes hard to see the error bars, the thin ones can be harder to track for those with poorer vision. Of course, color is not a necessity, but it can help to trace lines when, as in this example, they cross each other multiple times. Please also note that the size and type of symbols were changed - make them large enough and try to keep them consistent. There is a reason why, as in the panel taken from Prism, they are arranged in a certain order, it can sometimes guide you in your choice.
Your line must be thick enough to follow a curve clearly, but not so thick that it obscures other data points.
When using bar graphs you can choose patterns. But be careful, avoid overwhelming your reader.
When patterns don't fit, consider differentiating through color - since we live in 2025, you don’t need to worry that your figures will be printed in black and white.
Color Harmony
One major issue is that scientists believe they can choose colors arbitrarily.
But colors strongly influence how data is perceived. And without other supporting graphical elements, we cannot simply “overcome” biasing color.
And yes, figures like those on the left do get published… The only thing I want to add is that you should notice that the stroke width varies. The stroke on the right looks better because it is thick enough to be properly visible, yet remains within the same width range as the axes and the font, which makes the whole figure feel more integrated.
Rainbow color schemes, for example, draw attention unevenly, and distract the eye. Light tones such as bright yellows are more difficult to see.
There are dozens of palettes available online. The one shown above is from Color Hunt. Coolors also offers some with a larger range of colors. Of note, you don’t need to strictly adhere to a single palette; the key is to choose colors that look good together (and that don’t introduce visual bias).
Therefore, muted, darker tones are usually preferable.
If you want to use colors beyond grayscale, dark blues, purples, and - with some tact - orange or green tones are often good choices.
Next up, we just need to discuss graph types, statistics & composition.
What's a Society Journal? Hi Reader, let’s talk about something that has been essential to the development of the scientific system. They published the first scientific journal ever (despite it being far different from what we call a journal today). I am sure many scientists have even published in “them” without realizing it - I am referring to Society Journals. I bring this up because they might represent one of the best antidotes to predatory publishing and the larger issue of publication...
What's the Right Journal? Hi Reader, where do you normally publish your papers? But why exactly there? It remains one of the most important decisions for your career. If you’re unsure where to publish or if you’re considering switching journals, how do you find a good fit? Here is a pragmatic 7-step framework that should help you make a decision: Step 0 – Build an Initial List First, create a list of potential journals. While several will be top of mind, consider including those that you...
What Makes a Journal? Hi Reader, how many active, peer-reviewed academic journals exist today? There are 40,000! And are you interested in how many papers they publish every day? So, how can you know which journal to publish in? Today, we will discuss five key features of journals to help you differentiate them: What Differentiates Journals With so many journals available, it is easy to think that they differ only by name or impact factor. Click to enlarge. Please take these numbers with a...
