Develop Reference

Understanding of fig, ax, and plt when combining Matplotlib and Pandas

I'm trying to get a better understanding of how figure, axes, and plt all fit together when combining Matplotlib and Pandas for plotting. The accepted answer here helped me connect Matplotlib and Pandas in an object oriented way I understand through this line:
fig, ax = plt.suplots()
df.plot(ax=ax)
But as I'm diving deeper the answer here threw me off. Specifically, I still have methods I need to call directly off plt, that don't apply to either a figure or an axis. Example:
fig, ax = plt.subplots()
df[['realgdp','trend']]["2000-03-31":].plot(figsize=(8,8), ax=ax)
ax.set_title('Real GDP & Trend')
ax.set_ylabel('Readl GDP')
plt.xticks(rotation=45)
If I try to call xticks(rotation=45) off ax or fig I get an error that neither ax nor fig have an xticks method. The solution I have above works, but I don't understand why.
When I type plt.xticks(rotations=45), where does that information get sent? Why does the comment in the answer here that "when you use the functions available on the module pyplot you are plotting to the 'current figure' and 'current axes'" not apply in this case? Why do I need to call off plt directly?

plt.xticks() only works on the "current" ax. You should use ax.set_xticks(), ax.set_xticklabels() and ax.tick_params() instead.
plt.xticks() is a rather old function that is still supported, mimicking similar matlab code, born in a time when people were only plotting onto a single plot. The newer functions are more general with more options.
In short: you don't need to call plt directly, you are invited to use the ax functions instead. When calling plt.xticks(), it gets rerouted to the currently active ax (often the last one created).

Incomple text while saving seaborn figures [duplicate]

Updated MRE with subplots
I'm not sure of the usefulness of the original question and MRE. The margin padding seems to be properly adjusted for large x and y labels.
The issue is reproducible with subplots.
Using matplotlib 3.4.2
fig, axes = plt.subplots(ncols=2, nrows=2, figsize=(8, 6))
axes = axes.flatten()
for ax in axes:
ax.set_ylabel(r'$\ln\left(\frac{x_a-x_b}{x_a-x_c}\right)$')
ax.set_xlabel(r'$\ln\left(\frac{x_a-x_d}{x_a-x_e}\right)$')
plt.show()
Original
I am plotting a dataset using matplotlib where I have an xlabel that is quite "tall" (it's a formula rendered in TeX that contains a fraction and is therefore has the height equivalent of a couple of lines of text).
In any case, the bottom of the formula is always cut off when I draw the figures. Changing figure size doesn't seem to help this, and I haven't been able to figure out how to shift the x-axis "up" to make room for the xlabel. Something like that would be a reasonable temporary solution, but what would be nice would be to have a way to make matplotlib recognize automatically that the label is cut off and resize accordingly.
Here's an example of what I mean:
import matplotlib.pyplot as plt
plt.figure()
plt.ylabel(r'$\ln\left(\frac{x_a-x_b}{x_a-x_c}\right)$')
plt.xlabel(r'$\ln\left(\frac{x_a-x_d}{x_a-x_e}\right)$', fontsize=50)
plt.title('Example with matplotlib 3.4.2\nMRE no longer an issue')
plt.show()
The entire ylabel is visible, however, the xlabel is cut off at the bottom.
In the case this is a machine-specific problem, I am running this on OSX 10.6.8 with matplotlib 1.0.0

Use:
import matplotlib.pyplot as plt
plt.gcf().subplots_adjust(bottom=0.15)
# alternate option without .gcf
plt.subplots_adjust(bottom=0.15)
to make room for the label, where plt.gcf() means get the current figure. plt.gca(), which gets the current Axes, can also be used.
Edit:
Since I gave the answer, matplotlib has added the plt.tight_layout() function.
See matplotlib Tutorials: Tight Layout Guide
So I suggest using it:
fig, axes = plt.subplots(ncols=2, nrows=2, figsize=(8, 6))
axes = axes.flatten()
for ax in axes:
ax.set_ylabel(r'$\ln\left(\frac{x_a-x_b}{x_a-x_c}\right)$')
ax.set_xlabel(r'$\ln\left(\frac{x_a-x_d}{x_a-x_e}\right)$')
plt.tight_layout()
plt.show()

In case you want to store it to a file, you solve it using bbox_inches="tight" argument:
plt.savefig('myfile.png', bbox_inches="tight")

An easy option is to configure matplotlib to automatically adjust the plot size. It works perfectly for me and I'm not sure why it's not activated by default.
Method 1
Set this in your matplotlibrc file
figure.autolayout : True
See here for more information on customizing the matplotlibrc file: http://matplotlib.org/users/customizing.html
Method 2
Update the rcParams during runtime like this
from matplotlib import rcParams
rcParams.update({'figure.autolayout': True})
The advantage of using this approach is that your code will produce the same graphs on differently-configured machines.

plt.autoscale() worked for me.

You can also set custom padding as defaults in your $HOME/.matplotlib/matplotlib_rc as follows. In the example below I have modified both the bottom and left out-of-the-box padding:
# The figure subplot parameters. All dimensions are a fraction of the
# figure width or height
figure.subplot.left : 0.1 #left side of the subplots of the figure
#figure.subplot.right : 0.9
figure.subplot.bottom : 0.15
...

There is also a way to do this using the OOP interface, applying tight_layout directly to a figure:
fig, ax = plt.subplots()
fig.set_tight_layout(True)
https://matplotlib.org/stable/api/figure_api.html

for some reason sharex was set to True so I turned it back to False and it worked fine.
df.plot(........,sharex=False)

You need to use sizzors to modify the axis-range:
import sizzors as sizzors_module
sizzors_module.reshape_the_axis(plt).save("literlymylief.tiff")

Trying to add additional traces to an existing Matplotlib figure

I am using a function which spits out a figure object of validation data. My script calculates a few model parameters that I would like to plot on top of this existing figure object. How can I do this? Whenever I try to plot my modeled data, it does so in a new window. Here's what my code looks like:
datafig = plotting_function(args) #Returning a figure object
datafig.show()
plt.plot([modeled_x],[modeled_y]) #Plotting in a new window
I've tried using plt.hold() / plt.hold(True) but this doesn't do anything. Any ideas?
Edit:
MCVE:
import matplotlib.pyplot as plt
def fig_create():
fig_1, ax_1 = plt.subplots()
ax_1.plot([0,1],[0,1])
fig_2, ax_2 = plt.subplots()
ax_2.plot([0,1],[0,5])
return fig_1, ax_1, fig_2, ax_2
figure_1, axes_1, figure_2, axes_2 = fig_create()
plt.close("all") # Spyder plots even without a plt.show(), so running the function generates figures. I'm closing them here.
figure_2.show()
plt.figure(2)
plt.plot([0,1],[0,10])
Result of the MCVE: https://i.imgur.com/FiCJX33.png

You need to specify which axis to plot on. plt.figure(2) will make a figure with a number of 2, regardless of whether an existing figure has that number or not! axes_2.plot(), however will plot whatever data you input directly onto axes_2 and whatever was there already. If it doesn't immediately show up you should add plt.draw() after the plot function.
Try not to mix plt, notation and ax notation as this will create confusion later on! If you are using fig and ax, stick with those!

You can specify which figure to plot to by calling plt.figure(my_figure_index) before any plt.plot (or any other plt plotting function) call.
For example:
plt.figure(10) # creates new figure if doesn't exist yet
plt.plot(...) # plots in figure 10
plt.figure(2) # creates new figure if doesn't exist yet
plt.plot(...) # plots in this figure 2
plt.figure(10) # figure already exists, just makes it the active one
plt.plot(...) # plots in figure 10 (in addition to already existing stuff)

How modules know each other

I can plot data from a CSV file with the following code:
import pandas as pd
import matplotlib.pyplot as plt
df = pd.read_csv('test0.csv',delimiter='; ', engine='python')
df.plot(x='Column1', y='Column3')
plt.show()
But I don't understand one thing. How plt.show() knows about df? I'll make more sense to me seeing, somewhere, an expression like:
plt = something(df)
I have to mention I'm just learning Python.

Matplotlib has two "interfaces": a Matlab-style interface and an object-oriented interface.
Plotting with the Matlab-style interface looks like this:
import matplotlib.pyplot as plt
plt.plot(x, y)
plt.show()
The call to plt.plot implicitly creates a figure and an axes on which to draw.
The call to plt.show displays all figures.
Pandas is supporting the Matlab-style interface by implicitly creating a figure and axes for you when df.plot(x='Column1', y='Column3') is called.
Pandas can also use the more flexible object-oriented interface, in which case
your code would look like this:
import pandas as pd
import matplotlib.pyplot as plt
df = pd.read_csv('test0.csv',delimiter='; ', engine='python')
fig, ax = plt.subplots()
df.plot(ax=ax, x='Column1', y='Column3')
plt.show()
Here the axes, ax, is explicitly created and passed to df.plot, which then
calls ax.plot under the hood.
One case where the object-oriented interface is useful is when you wish to use
df.plot more than once while still drawing on the same axes:
fig, ax = plt.subplots()
df.plot(ax=ax, x='Column1', y='Column3')
df2.plot(ax=ax, x='Column2', y='Column4')
plt.show()

From the pandas docs on plotting:
The plot method on Series and DataFrame is just a simple wrapper
around :meth:plt.plot() <matplotlib.axes.Axes.plot>
So as is, the df.plot method is an highlevel call to plt.plot (using a wrapper), and thereafter, calling plt.show will simply:
display all figures and block until the figures have been closed
as it would with for all figures plotted with plt.plot.
Therefore, you don't see plt = something(df) as you would expect, because matpotlib.pyplot.plot is being called behind the scene by df.plot.

According to http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.show , the plt.show() itself doesn't know about the data, you need to pass the data as parameters.
What you are seeing should be the plot of pandas library, according to the usage http://pandas.pydata.org/pandas-docs/stable/visualization.html#basic-plotting-plot.
Hope this solves your question.

Why do pyplot methods apply instantly and subplot axes methods do not?

I'm editing my graphs step by step. Doing so, plt functions from matplotlib.pyplot apply instantly to my graphical output of pylab. That's great.
If I address axes of a subplot, it does not happen anymore.
Please find both alternatives in my minimal working example.
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
f = plt.figure()
sp1 = f.add_subplot(1,1,1)
f.show()
# This works well
sp1.set_xlim([1,5])
# Now I plot the graph
df = pd.Series([0,5,9,10,15])
df.hist(bins=50, color="red", alpha=0.5, normed=True, ax=sp1)
# ... and try to change the ticks of the x-axis
sp1.set_xticks(np.arange(1, 15, 1))
# Unfortunately, it does not result in an instant change
# because my plot has already been drawn.
# If I wanted to use the code above,
# I would have to execute him before drawing the graph.
# Therefore, I have to use this function:
plt.xticks(np.arange(1, 15, 1))
I understand that there is a difference between matplotlib.pyplot and an axis instance. Did I miss anything or does it just work this way?

Most of pyplot functions (if not all) have a call to plt.draw_if_interactive() before returning. So if you do
plt.ion()
plt.plot([1,2,3])
plt.xlim([-1,4])
you obtain that the plot is updated as you go. If you have interactive off, it won't create or update the plot until you don't call plt.show().
But all pyplot functions are wrappers around corresponding (usually) Axes methods.
If you want to use the OO interface, and still draw stuff as you type, you can do something like this
plt.ion() # if you don't have this, you probably don't get anything until you don't call a blocking `plt.show`
fig, ax = plt.subplots() # create an empty plot
ax.plot([1,2,3]) # create the line
plt.draw() # draw it (you can also use `draw_if_interactive`)
ax.set_xlim([-1,4]) #set the limits
plt.draw() # updata the plot
You don't have to use the pyplot you don't want, just remember to draw

The plt.xticks() method calls a function draw_if_interactive() that comes from pylab_setup(), who is updating the graph. In order to do it using sp1.set_xticks(), just call the corresponding show() method:
sp1.figure.show()