By pure impulse I decided to attend the data visualization workshop, something I definitely didn’t regret afterwards! Visualizing data in plots and figures is an important part of research – both so that colleagues don’t misunderstand our results in a research paper or presentation, but also so we don’t confuse ourselves during the research process. So how can we illustrate data in a simple and effective way without losing information?

The workshop was relaxed and interactive in nature. Many examples of good and bad data visualizations were given and the participants discussed how the plots and figures could be improved. Some brave participants also shared data of their own that they had a hard time presenting in a good way and left the workshop with many suggestions of improvements.

In summary, the main questions to focus on when visualizing data are

• What do you want to communicate with the figure?
• How can you highlight important features?
• How can you downplay less important features?

Information can always be encoded into a plot by adding for example colors, shapes or sizes to the data points. However, this should be done with caution as unwanted or unimportant features might accidentally become highlighted. For example, people are typically very bad at estimating how much bigger one area is compared to another. When encoding information by varying the area, as in a pie chart, it is likely that readers perceive differences in the data as larger than they actually are. Another common example is the heat map that uses more than two colors, classically going from blue to yellow to red. If the interest lies in what values are at the extremes of the heat map (red or blue), the yellow color will instead direct the attention towards the uninteresting middle values.

During this workshop, I learned many other useful tips and tricks that I will definitely keep in mind when making my next figure!

--Cathrine Bergh