The Python graph gallery mainly aims to help people with the technical part of data visualisation: it provides the code allowing to realise the chart you need. This section is a bit
different. It aims to list the most common mistakes of dataviz. Indeed, learning what a bad chart is is probably the most efficient way to avoid them. For each example, I try to provide
a few propositions to correct or improve the figure, always with python. Data visualisation is a huge field, and this section just shows a minuscule part of its complexity.
Note: this section is under construction. Drop me an email for any suggestion and please be indulgent!
A few resources list bad dataviz, not necessarily in python. It is a good idea to visit them if you want to improve your knowledge on data visualisation:
I am not a pie plot hater. There are a few rare case where they can be useful, and the general public is used to it.
However, it is definitely not the best way to represent data, and is often the theatre of awful dataviz. The gallery has a dedicated section if you still want to make one.
A Spaghetti plot is a line plot where many lines are displayed on the same chart. The chart becomes hard to read and thus do not deliver any insight. This post gives an examples, and propose several ways to avoid it.
Overplotting is a common problem in dataviz that occurs when the amount of data is huge. This post describes the problem (left chart) and proposes 10 solutions to avoid it, with reproducible code of course.
Boxplot and hidden data
A boxplot summarizes the distribution of a numerical variable for one or several groups. Thus, it hides the underlying distribution and the number of points of each group. That makes this chart dangerous. This post gives an example of possible mistake, and 3 solutions to fix it.