This was my second year running an interactive session at NWO Biophysics.
The topic: How to visualise a message by converting text into diagrams. The room was full with 35 PhD students and early-career researchers, all keen to find a way to translate a block of text into something visual and effective.
I have been into diagrams for a while now. I teach about them, run training sessions for researchers, and build them for grant-interview presentations.
Making a diagram seems simple; however, the hard part isn’t the design but everything that comes before it.
Texts are usually overloaded with information, and the hardest part is deciding what to keep and what to leave out. That’s a real struggle, because a diagram doesn’t tell the full story. It shows a relationship between elements, and it can only carry one message. Try to cram several in, and it gets just as messy and busy as the block of text it was meant to simplify.
The second issue is grouping. Once you know what to keep, how do you arrange it — by time, by category, by people? Choosing one can be harder than it sounds.
Then comes finding the relationship between the items.
If the first two steps were done properly, this part is usually straightforward, though spotting the pattern still takes some experience.
All three steps rest on the same thing: a clear message. Without one, you can’t tell which information matters, how to group it, or which relationship to highlight — every decision turns arbitrary. A clear message is what makes the rest of the process possible.
That’s also why elements in a text can often be related in more than one way, and I get asked the same question every time: which is the right diagram? My answer never changes: it depends on what your message is.
Take a CV. A timeline shows that you grew over time, progressing steadily or quickly in your career. A network shows the universities or organisations or teams with whom you have worked, to emphasise who you’d be bringing on board. A map shows the countries where you have worked, to emphasise experience across diverse cultural teams. Same information, three different messages.
Once these steps are clear, designing the diagram gets much easier. To help with that, I built a small widget: tell it how your elements are related, and it suggests the diagram type that shows that relationship best.
People also ask
What’s the most common mistake people make when turning text into a diagram?
Trying to include everything. A diagram isn’t a shorter version of the text — it’s one message, stripped down to the relationships that carry it, and that means genuinely cutting information, not just shrinking the font. Researchers are usually strong at spotting the pattern in their own data; they’re far less comfortable cutting the details that don’t serve it, because every detail took real work to produce. The skill worth building isn’t drawing boxes and arrows — it’s deciding, in advance, what the diagram is allowed to leave unsaid.
How do you choose what information to keep?
Think about the message you want to convey, then test each piece of information against it: does it support the message, or is it just a detail? A detail does not have to disappear — it can be said aloud during a presentation or added as a caption in a written piece. The diagram itself should carry one message only; everything else belongs in the voice-over, or in the surrounding text if it’s part of a proposal, paper or report.
How do you decide which diagram type fits your content?
It starts with the message and the relationship you want to emphasise, not the diagram. Once you know exactly what you want to say, the diagram type becomes a consequence of that message, not a style choice. The same set of facts can support several diagrams, and each one makes a different argument: a timeline says you progressed, a network says you’re well connected, a map says you have worked across cultures. None of them is more correct than the others — each one simply traces back to a different message.