I have recently started to appreciate the importance of illustrations and figures in scientific publications. Be it a visual abstract, a visual TOC, or part of a poster or a presentation, I always find a nice illustration to be quite catchy. Because of this, I want to put in some effort and learn what goes into preparing these illustrations.

Of course, I have the usual problem: I was never good in arts, and I can't really put my thought into a visual illustration, nor make the eventual idea into reality. Since I am working in theoretical physics I think I also have a harder time than some other sciences - the objects that I am working with are usually mathematical objects and proofs, and I can't really draw an equation, can I?

In contrast to illustrations, however, I've spent quite a lot of time on learning how to create good plots. In the case of plots I always had clear guidelines: from a given data, create a visualization that follows some basic principles with color scheming and choosing markers; using then the right software it's not hard to create visually appealing graphs. There's very little creativity involved in this process though.

Are there any similar "guidelines" for creating illustrations not based on qualitative data, rather abstract ideas? Is there maybe a book and/or a course that I might take?

I understand that scientific art and communication is a field in itself, however I am hoping that if I put in the time and effort, then, using the right tools, I could at least be mediocre in preparing illustrations. Or is creativity a pre-requisite for all of it?

  • 4
    While I don't agree with everything it says, this book is well worth a read.
    – avid
    Aug 23, 2021 at 19:01
  • Thank you for the recommendation; however, I am specifically looking for illustrations that are not based on data, but qualitative ideas. I am editing the post to make this more clear.
    – user145186
    Aug 23, 2021 at 19:09
  • 2
    'working in theoretical physics... I can't really draw an equation, can I?' Er, Feynman diagrams? Streamlines/field lines? Free body diagrams? Aug 24, 2021 at 11:43
  • @DanielHatton It is a nice suggestion, indeed some things can be drawn. But that doesn't mean I can come up a good illustration for a paper where I prove that my operator commutates with annother one under special circumstances.
    – user145186
    Aug 26, 2021 at 9:12
  • 1
    @LaBelleCroissant If your operator and the other one are expressed in Heisenberg matrix form for some system with a small finite number of eigenstates (e.g. spin up and spin down), can those matrices be interpreted as geometrical transformation operators (which, under your "special circumstances", produce the same result regardless of the order in which they're applied)? A geometrical transformation might lend itself to diagrammatic representation. Aug 26, 2021 at 10:12

4 Answers 4


For some fields, the (several) books by Edward Tufte set the standard. But a search of, say, Amazon for "Data Visualization" or "Scientific Illustration" will turn up a vast library of work. You can also add your field to the search to get a more specific selection.

But, like most things, the way to learn it is to practice and then practice more. Then try to get feedback on your efforts. Perhaps a colleague will tell you what the "see" in your work and you can compare the answers with your intent.


One of the great advantages of professional work is that it allows specialisation and gains from trade. One option you have is to hire an artist to create figures for you. There are a number of artists that specialise in scientific artworks, many of whom have a reasonable amount of scientific literacy to help them "tease out" how to visually represent scientific objects, concepts or ideas. Such work would usually require you to pay some money for services, and you would also give credit to the artist for the artwork in your paper, but it is likely to give you a much better illustration than if you create it yourself.


For scientific posters: There is a guy named Mike Morrison who has done a lot of work to make scientific posters better for the viewer. The concept is to make the main point (or take-home message) into a sentence or statement, and make this the focus of your poster. I've included an image to show this.

With standard posters, people oftentimes will read your title then ask you what the main point of your work is about. Having the main point front and centre tells the viewer this main point (without them having to ask), and then the viewer can ask you more in depth questions.

The benefits of using this design included:

  • Viewers that don't take the time to read through your poster will still know the main message.
  • Catches the audiences attention. They don't have to read through walls of text to find out what your main point is.
  • Less work for the reader, and so they can learn more at the conference.
  • Your time speaking with the viewer can be spent going into relevant details and engaging with the viewer, rather than getting to the main point.

I also recommend making sure your advisor is ok with using a format like this.

Image: This is an example poster. The main finding or take-home message is central to the poster. Additional information can be found on the side panels for you to reference when talking to someone, or for them to read on your own. Source: https://www.insidehighered.com/news/2019/06/24/theres-movement-better-scientific-posters-are-they-really-better

This is an example poster. The main finding or take-home message is central to the poster. Additional information can be found on the side panels for you to reference when talking to someone, or for them to read on your own. Source: https://www.insidehighered.com/news/2019/06/24/theres-movement-better-scientific-posters-are-they-really-better

References for further reading: Original: https://www.youtube.com/watch?v=SYk29tnxASs Updated: https://www.youtube.com/watch?v=1RwJbhkCA58 Article with some criticisms: https://www.insidehighered.com/news/2019/06/24/theres-movement-better-scientific-posters-are-they-really-better

  • 1
    What does Morrison recommend? We usually discourage link-only answers, because answers should stand on their own and links are subject to bit-rot. The approach sounds interesting enough to summarize it in an answer, though. Aug 24, 2021 at 14:41
  • 1
    Thanks for letting me know in a kind way. I didn't realize the etiquette. I have updated the post accordingly.
    – user1762
    Aug 24, 2021 at 15:04
  • While I personally find this particular example a bit too exaggerated, clearly emphasizing the key finding is certainly a good idea. It doesn't really answer the question though, which was about illustrations (possibly to be embedded in posters) and not posters per se.
    – Eike P.
    May 24, 2022 at 8:58

There are, I think, two different aspects to becoming a good scientific illustrator. The first, has to do with developing an ability to conceive the kind of illustration would be informative to a reader; the second has to do with developing the skills to bring the conception to fruition, either with traditional illustrative tools such as watercolor paints and oils (see, for example, this botanical illustration by Georg Dionysius Ehret, 1708-1770), lithography (see this superb example from Elizabeth Gould, 1804-1841), charcoal, etc. or by using modern computer driven approaches such as Inkscape and GIMP.

Of the two aspects, my personal experience has been that the problem of learning to use the tools (I can't draw and I can't paint, but I can use a mouse, a rollerball, or a sketchpad) pales in comparison with learning what it is that makes a good illustration and conceptualizing an appropriate illustration for a particular paper. Nonetheless, a good way to learn is to do what artists have done throughout the ages ... study other good illustrators. A source of constant amazement and inspiration to me are the illustrations in the magazine American Scientist (not to be confused with Scientific American). I'm not suggesting that every single illustration is noteworthy, but they are of consistently high quality and usually very appropriate for the article they serve. Additionally, many of the back copies of American Scientist are available, I believe, free of charge at JSTOR.

You must log in to answer this question.