Skip to content

Change Backend / Use figure toolbars

Use matplotlib figure toolbars in jupyter notebook environment.

1
matplotlib.use('nbagg')  # For Figure Tool bars embedded in jupyter output cell
Other backends:

1
2
3
matplotlib.use('TkAgg')
# or
matplotlib.use('Qt5Agg')
There are also few other cool backends built-in, which can be divided into 2 categories: Interactive and Non-interactive.

To see what's available:

1
2
3
from matplotlib.backends import backend_registry, BackendFilter
print("Interactives:", backend_registry.list_builtin(BackendFilter.INTERACTIVE))
print("Non-interactives:", backend_registry.list_builtin(BackendFilter.NON_INTERACTIVE))

Inline / commandline plotting

Inline + Interactive

While using ipython and a terminal that supports image rendering like kitty magic commands also work:

Technically this should work in jupyter notebooks too. 

1
2
3
%matplotlib tk
#or
%matplotlib qt

Kitty and matplotlib

Inline + Non-interactive

I am currently using kitty terminal. To use kitty as backend for python's matplotlib library. use the following:

  1. Install it:

    1
    pip install --user matplotlib-backend-kitty

  2. Use it (say in ipython or vanila python): 

    1
2
3
    import matplotlib
matplotlib.use('module://matplotlib-backend-kitty')
import matplotlib.pyplot as plt
    . More on this link.

Use Latex in global scope

It will display numbers in latex format and also the formula.

1
plt.rcParams['text.usetex'] = True

Format Dates in axis

Let's say I only want to show hours and minutes in the time axis (x-axis)

1
from matplotlib.dates import DateFormatter

1
2
3
time_format = DateFormatter("%H:%M")

ax.xaxis.set_major_formatter(time_format)

python date string formats (I like to call it python date format table) are found here