Machine Learning Udemy Bootcamp 05 - Matplotlib

It's the most used plotting library in data science and it was designed to have a similar feel to MatLab's (Math programming language) graphical plotting.
It works well with NumPy arrays and Pandas DataFrames.

conda install matplotlib
pip install matplotlib

Basics of Matplotlib

import matplotlib.pyplot as plt
import numpy as np
# if using Jupyter Notebook, include the following line to see plots in the notebook
%matplotlib inline

x = np.linspace(0,5,11)
y = x ** 2

Functional Method

plt.plot(
  # X axis values
  x,
  # Y axis values
  y,
  # function red line
  'r-'
)
# Define a label for X axis
plt.xlabel('X Label')
# Define a label for Y axis
plt.ylabel('Y Label')
# Define a title for the plot
plt.title('Title')
# display the plot programattically (outside jupyer notebook)
plt.show()


# Plot multiple plots on the same canvas
plt.subplot(1,2,1)
plt.plot(x,y,'r')
plt.subplot(1,2,2)
plt.plot(y,x,'b')

OOP Method

A basic Matplotlib figure is composed by two objects: a figure and axes. The figure is the canvas where the plot is drawn and the axes are the plot itself.

# Create a figure object, imagine it as a blank canvas
fig = plt.figure()

axes = fig.add_axes([
    # axis left
    0.1,
    # axis bottom
    0.1,
    # axis width
    0.8,
    # axis height
    0.8
])

axes.plot(x,y)
axes.set_xlabel('X Label')
axes.set_ylabel('Y Label')
axes.set_title('Title')

Multiple plots on the same canvas:

fig = plt.figure()
# Create 2 sets of axes on the same canvas
axes1 = fig.add_axes([0.1,0.1,0.8,0.8])
axes2 = fig.add_axes([
  # 20% from the left
  0.2,
  # 50% from the bottom
  0.5,
  # 40% of canvas total width
  0.4,
  # 30% of canvas total height
  0.3
])

# function
axes1.plot(x,y)
axes1.set_title('Larger Plot')
# reverse function
axes2.plot(y,x)
axes2.set_title('Smaller Plot')

Shorthand to create a figure with axes in OOP:

# automatically invoke fig.add_axes based on nrows and ncols params
# this creates a grid 1x2 containing a plot on each cell (same as with 3x3 grid)
fig, axes = plt.subplots(
  # number of rows, positional or keyworded
  nrows=1,
  # number of columns, positional or keyworded
  ncols=2
)

# improves the layout of the plots, prevents overlapping
plt.tight_layout()

# axes is a list of matplotlib axes, we're able to iterate over it
for axis in axes:
  # set same x and y values for all plots in the canvas
  axis.plot(x,y)

# manually iterate over axes
axes[0].plot(x,y)
axes[0].set_title('First Plot')
axes[1].plot(y,x)
axes[1].set_title('Second Plot')

Figure Size, Aspect Ratio and DPI

fig = plt.figure(
    # figsize is a tuple with width and height in inches
    figsize=(8,2),
    # dpi is the dots per inch (pixel per inch)
    dpi=100
)

ax = fig.add_axes([0,0,1,1])
ax.plot(x,y)

Shorthand:

fig, axes = plt.subplots(nrows=2, ncols=1, figsize=(8,2))
axes[0].plot(x,y)
axes[1].plot(y,x)
plt.tight_layout()

fig.savefig('my_picture.png', dpi=200)

Multi functions on same plot:

fig = plt.figure()
ax = fig.add_axes([0,0,1,1])
ax.plot(x,x**2, label='X Squared')
ax.plot(x,x**3, label='X Cubed')

# show a leger the matches labes with functions color
ax.legend(
  # value 0 to 10 to define the position of the legend
  # using 0 matplotlib will understand what's the best position 
  loc=0
)

Plot Appearance

fig = plt.figure()
ax = fig.add_axes([0,0,1,1])
ax.plot(
  # data on x
  x,
  # data on y
  y,
  # color of the line as name or hex or rgb
  color='purple',
  # line width, default is 1. Also works as lw=3
  linewidth=3,
  # opacity of the line
  alpha=0.5,
  # linestyle, - is solid, -- is dashed, -. is dashdot, : is dotted, steps is disquared. Also works as ls='-'
  linestyle='-',
  # marker, o is circle, + is plus, * is star, s is square, 1 is triangle down, 2 is triangle up, 3 is triangle left, 4 is triangle right, x is x, | is vertical line, _ is horizontal line. Also works as marker='o'
  # appearence of dots on the line
  marker='o',
  # size of the marker
  markersize=20,
  # marker color
  markerfacecolor='yellow',
  # marker border width
  markeredgewidth=3,
  # marker border color
  markeredgecolor='green'
)

Plot limits

fig = plt.figure()
ax = fig.add_axes([0,0,1,1])
ax.plot(x,y,color='purple',linewidth=2,alpha=0.5,linestyle='--',marker='o',markersize=20,markerfacecolor='yellow',markeredgewidth=3,markeredgecolor='green')
# distribute X axis values from 0 to 1
ax.set_xlim([0,1])
# distribute Y axis values from 0 to 1
ax.set_ylim([0,2])