GridSpec for dummies

When we want to create a figure that includes more than one plot, while having the possibility to manage each of these individually, we can normally achieve this by using the pyplot.subplots command. However, there are complex cases that pyplot.subplots cannot handle. In these cases, we should use the GridSpec function, also from matplotlib.

It can be difficult to generate the desired grid and trying to allocate each of the plot to its correct coordinates, specially, when the plot spans multiple rows or columns. I recommend to first draw a draft of the desired plot layout. And then divide the figure in rows (red lines) and columns (blue lines) as the example. Like this, you know the particular coordinates for a plot. Then, the coding is really easy.

#Import necessary libraries
import matplotlib.pyplot as plt
from matplotlib.gridspec import GridSpec

#Initialise figure
fig = plt.figure(layout = "constrained")

#Initialise Gridspec (number of rows, number of columns, figure)
gs = GridSpec(4, 2, figure = fig)

#Add grids to add the plots later
#Plot A spans row 0 and column 0
ax_a = fig.add_subplot(gs[0, 0]) 
ax_b = fig.add_subplot(gs[0, 1])
#Plot C spans rows 1 and 2; and column 0
ax_c = fig.add_subplot(gs[1:3, 0]) #Remember the slicing rules in Python
ax_d = fig.add_subplot(gs[1, 1])
ax_e = fig.add_subplot(gs[2, 1])
#Plot D spans row 3 and all the columns
ax_f = fig.add_subplot(gs[3, 0:])

Author: Olivia Dove Estrella