: How do I improve the visual design of my scientific schematic figures? When writing a manuscript, I generally feel like my plots follow a good balance of visibility and informational portrayal.

@Hamaas979

Whether you like it or not, a lot of design choices will be dictated to you by the constraints of the medium:


You have to expect that a relevant amount of the readers will print your paper on a typical monochrome laser printer, which should largely predetermine your colour choices:


You have to expect that there are only four colours that are reasonably distinguishable in the printout: black, light grey, dark grey and black. Any other colour will either default to one of these colours, be indistinguishable from them or have undefined behaviour. Therefore, the only safe thing is to resort to these four colours (or colours who are mapped onto them by greyscale conversion).
You have to expect dark grey texts to be more difficult to read and anything lighter to be unreadable. You have to expect lines in light grey to be invisible.
You have to expect that the exact position of borders between areas coloured anything other than black and white cannot be clearly made out. If this is relevant, draw a black or white line in between.
Non-greyscale colours can only be used as a bonus to increase differentiability for those viewing the figure on screen or in colour print, but you cannot rely on them.
Dark backgrounds will not work out.
Gradients will not work out.

Your figures are expected not to waste much space. Therefore you cannot leave large parts of the figure empty and in particular you have to expect that the figure will be cropped to its relevant contents in printout. The printing department might even butcher your figure to make it smaller.
Some publishers have instructions regarding which font to use in figures. As these fonts are usually only available to users of a certain operating system, I assume that this is mainly to avoid people using Comic Sans and similar – but you never know. It is not unlikely though that some publisher is adamant regarding serif and sans-serif.


Regarding your example figure, I noticed the following things (mind though, that I am not a professional graphics designer):


If I assume this figure to be scaled to column width in a two-column layout, the lines are rather thin.
Consider using completely black shapes and white shapes with black border only. (But this really depends on what the figure is about.)
If the light grey fill will not be practically white in print, it has a good chance to look ugly or to be (slightly) confusing as it only differs from the background at third glance. Therefore either make it darker or plain white.
One of the circles is closer to the pill than stroke width. Unless this one is special, it unnecessarily attracts the attention of the viewer.
The pill’s diameter is slightly larger than the circles’ one. This unnecessarily irritates the viewer.


Finally, some general advice that does not apply to your example:


Avoid using more than one font size.
Avoid using more than one stroke width.
Align components to each other if necessary.
Prefer simple geometric shapes and angles.
Use vector graphics.

Vote Up Vote Down

@Steve758

Well, not a really easy question since there are no "correct answers".

First, I would stay away from using maxed black colors (that is a color that appears to be 100% black, 100/100/100/100 in CMYK for example) since that looks unnatural. When you see black in real life, it's not complete black - it's a shade. This applies to texts as well as to outlines. I used 80% K in CMYK on the outlines and the texts.

Use subtle gradients on the background and on each figure to make them look a bit friendlier, and also more natural. I used round gradients in my example.

Use standard sans serif fonts, like Helvetica, as a safe bet for typing.

Vote Up Vote Down