10 Results and visualisation
UNDER CONSTRUCTION
10.1 Plotting and scales
10.1.1 Scales and white space
Some white space looks sophisticated, too much and it looks like you don’t have any results.
So scale your axes to show the differences, but there’s a tension with not being misleading.
10.1.2 Font and symbol size
The default for styling plots should be for easy viewing on mobile devices.
Most digital interactions are via mobile these days. Even if you plan on viewing your plots on a desktop, a mobile friendly layout will make them more readable.
I recommend defaulting to a mobile friendly view from day 1 of data exploration. This means using larger font, line and label sizes than standard plotting package defaults. I’ll show how below.
You can then alter the styling as your finalize the plots of publication, where desktop viewing might be more common (I saw ‘might be’ because I frequently papers on my phone these days).
Styling plots for visual clarity is key from the start, because clarity of presentation influences how you interpret the research. So it has a tangible impact on the way your research will develop and how you engage with your collaborators.
I often communicate with collaborators and students via instant messages (e.g. Teams), which allows for quick feedback cycles. The default settings can be hard to view however. Larger font is also easier for a broader audience to read.
If you are using R and ggplot2 for plots, then this one-liner will make the font on all your plots bigger theme_set(theme_bw(base_size = 28)).
10.1.3 Fancy and complex plots
If its getting too complex, think about what you are trying to communicate, then split your plot into several plots, one for each point.