I recently received feedback for a university-intern seminar paper where the reviewer implicitly stated, that every academic paper needs graphics. The paper was in the field of computer science.

I am aware, that graphics can contribute towards understanding an issue, however, for this specific topic I stood before the issue that

  • Graphics I would have liked to use are copyright-protected, thus no option
  • I am a bad drawer
  • There seemed no fitting option to add computer-aided drawings such like diagrams (in terms of content, I do know how to use Latex)

Therefore, I concluded that I would not include any graphics in my work.

I have encountered several papers on the internet, that also included no graphics. Plus, naturally, mathematical papers would often also not need graphics depending on the topic, with computer science being an applied case of mathematics.

Are graphics in academic papers a real requirement or should I consider the feedback as a subjective need?

  • 6
    I would say that it's actually pretty rare that a paper can't benefit from a diagram that illustrates the intuition of a theorem, or a plot that shows the results of an experiment more directly than a table could. Such papers exist but they're quite rare. Most papers without some graphics could probably be improved with them - but page count limits in a journal might not have permitted it. Squeezing in graphics often means squeezing out text.
    – ObscureOwl
    Commented Jan 19, 2021 at 22:10
  • Few points: 1) there are many open source cliparts. Also, you are free to redraw diagrams and cite where you adopted a diagram, 2) bad drawer? It does not have to be at a high level. Block diagrams, flowcharts, etc.. simple diagrams will do. 3) not sure I understand your third point. What do you mean by 'to add'? Unless you are using text based only editors, MS Word and LaTeX allow one to easily include graphics. Commented Jan 19, 2021 at 22:13
  • 2
    It's a different setting, but some of the answers to Referee wants me to add a plot – because the paper has none may be relevant.
    – Anyon
    Commented Jan 19, 2021 at 22:31
  • I rephrased the sentence. I do know how to add graphics, and I do know how to create diagrams, but there seemed to be no sense in artificially crafting some with low-quality content just to have some.
    – kaiya
    Commented Jan 19, 2021 at 22:39
  • 1
    "I would say that it's actually pretty rare that a paper can't benefit from a diagram" - That's @ObscureOwl not the point: If it were for that, no paper would be ever accepted. There is always things one can add which profit a paper, at least for part of the readers.
    – user151413
    Commented Jan 20, 2021 at 0:07

4 Answers 4


In (sufficiently) theoretical fields, such as theoretical computer science, parts of theoretical physics, or mathematics, it is completely fine to publish a paper without graphics.

Whether in the case of your specific paper, topic, or field graphics would have helped is a different question, and essentially impossible to answer here with the information given. On the other hand, if it is a field where papers with no graphics are reasonably common, there should be not a big issue in convincing the editor that graphics are unnecessary.

Note that similar arguments could be applied to other things, like examples, applications, etc.: Nice to have, they likely don't hurt, but not needed in sufficiently abstract fields of research.


I think the reviewer was overstating the need. They are often helpful and sometimes essential, but not in every paper as you have seen.

But you might want to explore drawing programs more fully. And most composing systems from MS Word on up permit you to add graphics, sometimes by just cut and paste. Most graphics programs will create standard file format outputs such as png and jpg files that are acceptable in a lot of work.

For a school project, however, you need to do what is required, even if it is more restricted than normal.

Graphics are needed when they are needed, but not otherwise.

  • I rephrased the question. I do know how to use Latex and I am able to draw diagrams and create plots etc with R. Its just, there is no use case for that in this paper.
    – kaiya
    Commented Jan 19, 2021 at 22:40
  • 1
    Good. I'll leave it as is for the possible benefit of future readers.
    – Buffy
    Commented Jan 19, 2021 at 22:43
  • Please don't use raster formats for diagrams, graphs etc., especially lossily compressed ones like JPG. Use vector formats like SVG or PDF, and save raster formats for photographs and similar
    – Chris H
    Commented Jan 20, 2021 at 15:16

If you want to be a scientist (and I suspect this applies to Computer Science) then you must learn to do some graphic design, or have enough money to pay someone to do it for you.

Admittedly there are some fields of research where nobody uses graphics. But in the fields where I work, >90% of papers have 3+ graphics. One of the reasons for the graphics is that many of your casual readers are not willing to read your entire paper, but they are willing to look at nice-looking graphics. If you can make graphics convey the point of your paper, then you will communicate with those casual readers.

Graphics I would have liked to use are copyright-protected, thus no option

It is customary to create your own graphics, even if you have a right to use other people's graphics.

No option to add computer-aided drawings like diagrams

That is an option. For diagrams, I suggest you learn TikZ.


  • I have already used TikZ and yEd for diagrams. I am aware, that own graphics are preferred. But for the recent topic it simply makes no sense to draw diagrams. All I picture someone could do would be to redraw images of the described hardware, however, when enumerating multiple of them it seems not possible to add a graphic for each of them due to space limitations and it seems totally arbitrary and unprofessional to add graphics for only some of them. And these pictures would have no value at all. I do not think that for the topic I chose, any graphic could convey the point of my paper...
    – kaiya
    Commented Jan 19, 2021 at 22:36
  • 1
    Ask your supervisor. Commented Jan 19, 2021 at 22:56
  • Of course. But in addition to the context-related answer, I also want to have the global answer, thus the question here.
    – kaiya
    Commented Jan 19, 2021 at 23:00
  • 1
    "readers are not willing to read your entire paper, but they are willing to look at nice-looking graphics" - Every paper gets the readers it deserves.
    – user151413
    Commented Jan 20, 2021 at 0:08

Similar to Anonymous Physicist, I think graphics are important, although I think "casual reader" is not the right viewpoint.

Academics have to read lots of papers. But there are many more papers than you really have time to read. You have to learn a method for evaluating whether a given paper is really worth spending serious time to read. A good example is given here by Andrew Ng. As he says, just reading a paper from first word to last is a bad way to do it. Rather, you read in multiple passes, and you can decide halfway through whether it's worth continuing. What do you read in the first pass? The title, abstract, and you look at the figures. (And other "how to read an academic paper" lectures tend to make the same points.)

So this isn't really a case of readers being shallow. Rather, of readers having time that they could be spending reading a different paper, if they can't quickly see whether your paper is valuable to them.

You can do without graphics, but you'd better have an amazing abstract.

  • Thanks for the answer. I think, I have a good point why people should read the paper, that is clearly visible when skimming the paper (due to formatting choices), plus the topic rare enough, that interested researchers are likely required to read all applicable papers they come across, anyway, as long as they seem to provide new insights. I do get your point with images, though and appreciate your explanation.
    – kaiya
    Commented Jan 20, 2021 at 13:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .