Skip to content

2. Story

A figure can express an idea quickly, succinctly, and straightforwardly. However, it is important to know exactly what the message of the figure is, to avoid creating either pointless plots that don’t really tell much, or overly complex plots that try to tell far too much.

Diagram first

Try sketching out some ideas with pen and paper to explore how you might want to present the data. Exploratory data vis (discussed in the previous slide) is also useful for this. - What's the result you want to showcase? Can you explain it in a single sentence? - Do you need to show the broad trends or the fine-grained details to tell the story? - These should usually be separated; if both are relevant, split them into two separate plots - What is the simplest graph possible that will communicate the meaning?

Collaborate and share

Share your plots from an early stage (perhaps even at the sketching-with-pen-and-paper stage) to see if your peers can quickly grasp the message you are hoping to convey.

You know your data and your research very well; the patterns you are trying to highlight might be very obvious to you when you look at your figure, but completely hidden to someone looking at it for the first time.

What sort of plots are even out there?

When trying to figure out what kind of plot wll best help to visualise your message, it's useful to remind yourself of what kind of plots even exist!

Places to look for plotting information

Plotting library galleries and showcases

A great place to find out whats available and what you can achieve is to scroll through some of the well-curated example galleries and tutorials published alongside many of the popular Python plotting libraries.

Some notable examples include:

There are also some larger-scale collected galleries that include multiple libraries: - The Python Graph Gallery - This gallery breaks up plots from a range of different libraries into topics such as distribution, correlation, ranking, evolution, and mapping

Articles on data visualisation

While not specifically sharing Python libraries, general articles discussing great data visualisation can be a useful place to discover new kinds of plots and to inspire yourself; these sorts of posts can often be found on company blogs

Research articles

When keeping up-to-date on research in your field, make sure to save any impressive figures that you want to draw inspiration from at a later point!

Some of the high-budget, flashier journals often have impressive graphics in their featured articles (for example, Nature). Look outside of your research area to discover new ways of plotting.