Humans are often better at interpreting messages graphically compared to numerically. It is therefor important to create good figures. Plots are useful for summarization and showing trends in data.

Creating good figures

Here are some things to consider when creating a plot:

Be sparing: When choosing to employ visualizations, they should support the messsage and aid in communication.
Keep It Simple Stupid (KISS): It is often better with a simple plot rather than a cluttered one.
Minimize unnecessary elements and decoration
- Some plotting tools may include unnecessary and disturbing elements by default.
- Gridlines, boxes, secondary ticks, etc. may be unnecessary sometimes.
Use the same fonts as in the text that the plot will be placed in (Check out LaTeX mode in matplotlib)
Produce vector graphics (e.g. PDF) rather than a PNG to keep plot looking sharp even when zooming in.

Lie Factor

Minimizing the lie factor is important to avoid misleading plots.

$\text{Lie Factor} = \frac{\text{size of an effect in the graphic}}{\text{size of the effect in the data}}$
Present the actual data, not just interpolations.
Use appropriate scale for the figure
Show tick labels
Show the origin
Do not use 3D charts or pie charts, they are hard to interpret.

Data visualization

Creating good figures

Lie Factor