Matplotlib: How to change figsize for double bar plot

Question

I have plotted a double bar plot in matplotlib using the following code:

x = pd.Series(range(12))
y = self.cust_data['Cluster_ID'].value_counts().sort_index()
z = self.cust_data['Cluster_ID_NEW'].value_counts().sort_index()
plt.bar(x + 0.0, y, color = 'b', width = 0.5)
plt.bar(x + 0.5, z, color = 'g', width = 0.5)
plt.legend(['Old Cluster', 'New Cluster'])
plt.savefig("C:\\Users\\utkarsh.a.ranjan\\Documents\\pyqt_data\\generate_results\\bar", bbox_inches='tight',pad_inches=0.1)
plt.clf()

I want to use the figsize parameter to make the resultant plot bigger in size. This is easy when plotting a single bar plot, but here I am confused as to where to put the figsize parameter.

Answer 1

The best way to do that follows a more OO approach:

fig, ax = plt.subplots(figsize=(12,12))
ax.bar(x + 0.0, y, color = 'b', width = 0.5)
ax.bar(x + 0.5, z, color = 'g', width = 0.5)
ax.legend(['Old Cluster', 'New Cluster'])
fig.savefig("C:\\Users\\utkarsh.a.ranjan\\Documents\\pyqt_data\\generate_results\\bar", bbox_inches='tight',pad_inches=0.1)
plt.clf()

You also might want to add format to your filename. If not, the format is taken from your rc parameter savefig.format, which is usually png.

BTW, if you want to stick to your code as much as possible, you can also add the line before your plt.bar(...):

plt.figure(figsize=(12,12))

Answer 2

A very simple way to increase plot size is to simply add this line of code before your code:

plt.rcParams["figure.figsize"] = (10, 5)

You can change the figure size according to your wish.

Answer 3

You could set the size with figsize

import matplotlib.pyplot as plt

f, ax = plt.subplots(figsize=(18,5)) # set the size that you'd like (width, height)
plt.bar([1,2,3,4], [0.1,0.2,0.3,0.4], label = 'first bar')
plt.bar([10,11,12,13], [0.4,0.3,0.2,0.1], label = 'second bar')
ax.legend(fontsize = 14)