Select Page

Have you ever looked at a plot and wished you could understand its point? If data is so important to support our science, plot design should be an essential part of your writing.

Example of data plot design

In a Master thesis, I supervised, I helped my student with a plot (see figure above) to express where our data was relative to other authors, and also include a subplot to show a more detailed view of our data. One of the Juri members told that plot said everything.

Edward Tufte is one of the world experts in the visualization of data. He has a sentence I use in my classes on Programming when students have to program to perform plot design from data. He says in his book on Visual Explanations that “… clarity and excellence in thinking is very much like clarity and excellence in the display of data.”  I really like this sentence because it expresses exactly how I feel about a plot.

Good plots mean good thinking.


What tools

When I used Windows my favorite tool was Origin. However, when I changed to the macOS ecosystem, it was a bit difficult to find the right tool. All were relatively expensive and not found at my University until I crossed path with DataGraph.

The app has a good price and soon became my main plotting tool. But I experienced something else over the years. A single plot tool is not enough. The final art often involves other apps.

For example, to include equations I use LaTeXiT, and to make final art adjustment I use Autodesk’s Graphic. But this is the technical part.


Design options

Everyone has his own taste. Because plotting contains a bit of art, what I think is a good way of plotting may differ from what you think. However, there are a few reasons for making certain design decisions.

Once I heard a plot should be read in the diagonal, meaning all data is between the lower left and upper right corners. This is a good guideline, unless you’re comparing data need to adjust the axis range to include the dispersion of values in different plots.

Then, I always point to

  • large symbols;
  • thick axis;
  • large axis labels and larger and bold axis titles;
  • and boxed plots.

It could be simple plot design options, but the reason is to improve, as much as possible, the reading of the plot when we insert it into a paper or document. The plot will shrink, and if its elements are not large or thick enough, it becomes rather small and difficult to visualize.


Plot design options

If you’re using LaTeX to write your paper, you might export your plot as PDF files because the output is vectorial. This means it doesn’t lose resolution when you adjust the size of the plot in your document. Otherwise, if the plot contains “heavy” elements, like pictures or a significant number of data points, or even 3D rendering, you might consider exporting as a PNG (Portable Network Graphic) instead.

Finally, in my opinion, when I design a plot, I always aim for self-explanatory graphics. The reader shouldn’t need to read the text to understand what the plot shows. Sometimes this is not possible, other times it’s not. The important is to make simple, clear and visually appealing, like presentations, in such a way that it expresses the clarity and excellence in thinking.


QUESTION: what tool do you prefer in your plot design?