Scientific writing 3.0: A reader and writer's guide - Jean-Luc Lebrun, Justin Lebrun 2021

Visuals Q&A
Visuals: the voice of your paper
Paper structure and purpose

Q: I would like to recycle a visual I first published in the extended abstract of a conference proceeding to reuse it for a journal publication. The two papers are very close. Can I just reuse it or do I need to redo it?

A: Unless you have kept the copyright to the original paper, you are obligated to ask the owner of the conference proceedings for the permission to re-use the visual. If the two papers are very close, you also have to give the journal a copy of the original conference paper so that the editor can assert whether there are enough differences to merit publication. In short, redo the visual. There are many ways to present data, and I am sure that the original visual could benefit from another draft. After all, any paragraph published, or any visual published is only a draft. There are so many ways to improve on text, likewise for visuals.

Q: Can I change the contrast of my visual to make it more readable?

A: Any image post-treatment is always looked upon with suspicion, and is often forbidden by the journal, so check the journal’s instructions to its authors regarding artwork, and abide by them. There is some tolerance, for example to improve contrast, but again, 1) if improving contrast removes information from your visual, or conceals it, you should not do it (increasing the contrast may remove a gray level that was significant); and 2) you should inform the journal that you have manipulated the image, and provide the original image for comparison.

Q: Can a reference to a figure (say figure 2) be made in the caption of another figure (say figure 3)?

A: Preferably not. Figure 3 is not standalone because understanding it requires the prior understanding of figure 2. However if you cannot make figure 3 stand-alone, do mention figure 2 in its caption. It will at least guide the reader to find all relevant information.

Q: Can I write the interpretation of the visual in its caption, or do I leave that to the discussion?

A: If the visual is in the result section of your paper, you are not expected to interpret it there. But remember, the way you have laid out your data already guides the reader towards your interpretation. Highlight the salient data or points that will be used in the discussion section when you offer your interpretation. If the visual is in the discussion section, nothing should stop you from providing an interpretation in the caption. This will definitely help you achieve the worthwhile goal of making any visual stand-alone.

Q: Do I need to bother about page layout issues for my visuals since the journal is going to redraw them anyway?

A: You should if you want to increase the chances for your table or graph to be kept close to its reference in the text. This is particularly relevant when the journal you are targeting has multiple columns on one page. When that is the case, you may want to redesign a horizontal table to make it fit vertically on one column of a two column page. For example, Visual 11 in this chapter could easily be made vertical by inverting the x-y-axis with no loss of meaning! But keep vertical what people expect to see vertical, like temperature bars; and keep horizontal what which people expect to see horizontal, like distances. And while on the subject of visuals, never hesitate to insert additional space or lines to add structure and to highlight what you consider important. Use (larger) sans serif type fonts if you expect that the visual will have to be shrunk to fit the size of a page or column. Better still; rework your visual to make it fit on a page or column naturally by increasing conciseness and clarity.

Q: There are so many ways to visualize my data, so which way is the best?

A: Before you even ask yourself that question, ask yourself these, IN THE ORDER GIVEN:

· Which point do I want to make in my visual? Asking this question helps you avoid the pointless data dump.

· What data do I absolutely need to make my point, and which data can I leave out?

· Do I have the data I need to make the point I want to make? If not, what point can I make with the data I have?

· Which level of data aggregation or transformation (frequency, percentage, mean, cumulative, log, etc) will best make the point?

· Now that I have the right data, and I know what point to make with it, what is its dimensionality (2D, 3D, nD vector)? What is its nature (qualitative, discrete, continuous, temporal, pictorial, numerical, symbolic as a chemical notation or the elements in a diagram)? What is its accuracy (error bar, range, probability, resolution)? What is its range and scale (finite, infinite, zero to one, set of known attributes, set of given names, etc)?

· Given my type of data, which ordering scheme would support my point: the data intrinsic order (chronological, numerical, spatial, logical, hierarchical), or a new order such as functional proximity, functional classes, inclusion — exclusion, with — without, before — after, generic — specific, simple — complex, most probable — least probable, low priority — high priority, most favorable — less favorable, most relevant — less relevant, similar — different, parents — children, etc)?

· Which visual representation is the one the reader expects to see in order to be convinced by my ordered data? (more than one may apply — use the one which makes the point most immediately and most clearly) Table, list, line chart, stacked line chart, flowchart, bar graph, photo, Venn diagram, block diagram, tree structure, schematic diagram, 2D bar chart, 3D line chart, Pie chart, doughnut chart, 2D area chart, 3D floating bars, point chart, bubble chart, grid surface chart, mesh surface chart, radar chart, polar chart, high-low chart, error bar chart, funnel chart, box chart, XYY chart, Western blot, Northern blot, Pareto chart, scatter plot, and so many other representations are available.

Q: Why do I get lost when I look at some visuals, and what should I do so that my readers do not get lost in my visuals?

A: Contrary to the paragraph, which has only one point of entry (its first words), a visual can have many points of entry. Unguided, the reader’s eyes dart here and there, trying to find what there is to learn from that visual. You must design the visual such that the eyes are guided in their exploration through the visual. The title of a figure or table serves the same role as the topic sentence in a paragraph. It helps the reader choose the right entry points. The photo of enzyme entrapment in siliceous foam could be titled “Evidence of Enzyme nitrogen (b) trapped by pressure in siliceous foam (a)”. Depending on the point you want to make, the title of the one-two step comparison table could either be “The higher the one step positive rate, the higher the rate increase from the second step PSY”, or “Adding a second step is not computationally efficient except for the MSV-PSY pairing”. While visuals allow for free visual exploration, these titles guide the reader’s eyes onto a specific path important to you, the writer.

Q: How do I guide the reader towards the main visual in my paper?

A: Make it stand out. You could make that visual the largest one, or the only one with color. You could put many words from the title of your paper in the visual’s title or in in the first line of its caption, or you could simply start the caption with “This figure represents the core of our contribution.”

Q: Should I repeat what the X and Y axis mean in the caption?

A: There is no need to repeat whatever is blatantly obvious to all in the figure itself. Write instead on whatever is NOT obvious and is necessary to make that visual understandable and stand-alone. Do not repeat the figure; explain it.