visualizations.md

Visualizations with matplotlib.pyplot

All of the following code will only work in a gui environment. In other words, if you are connected to a remote machine through the terminal (pythonanywhere, kaggle, Google Colab, ssh, etc.), that machine has no way to open a new window to show you the generated image. However, you can save image files by using the following idiom instead of plt.show():

>>> fig = plt.gcf()  # get current figure
>>> fig.savefig('filename.png')
>>> plt.clf()  # clear plot (only necessary if you need to make more images)

The following code is adapted from this website: https://matplotlib.org/tutorials/introductory/pyplot.html#sphx-glr-tutorials-introductory-pyplot-py

drawing simple line

import matplotlib.pyplot as plt  # $ python3 -m pip install --user matplotlib
seq = [1, 2, 3, 4, 5, 6]
plt.plot(seq)  # x is assumed to be range(len(seq)), y is seq
plt.ylabel('some numbers')
plt.show()

drawing parabolic line

import matplotlib.pyplot as plt
seq2 = [i ** 2 for i in seq]
plt.plot(seq, seq2)  # seq is x, seq2 is y
plt.show()

PRACTICE A

Write a new list myseq and try generating plots. Change the numbers to higher and lower values. Use more or fewer numbers. Play around until you are comfortable drawing line graphs.

Plotting points using custom marking shape/line/color

For all marker formatting options, see https://matplotlib.org/api/_as_gen/matplotlib.pyplot.plot.html#matplotlib.pyplot.plot

import matplotlib.pyplot as plt
plt.plot(seq, seq2, 'ro')  # r = red; o = circle
plt.axis([0, 10, 0, 100])  # Usually automatic axis lengths are best; this is just an example of how to change it if needed.
plt.show()

PRACTICE B

Use different combinations of marker formatting on the website shown above. Draw lines and dots of different colors and shapes. Tinker.

Drawing multiple series using various shapes/colors

You can plot multiples series all at once...

import matplotlib.pyplot as plt
seq3 = [i ** 3 for i in seq]
plt.plot(seq, seq, 'r--', seq, seq2, 'bs', seq, seq3, 'g^')
plt.show()

...or one at a time, before calling plt.show()... (I think this is more readable)

import matplotlib.pyplot as plt
seq = [1, 2, 3, 4, 5, 6]
seq2 = [i ** 2 for i in seq]
seq3 = [i ** 3 for i in seq]
plt.plot(seq, seq, 'r--')
plt.plot(seq, seq2, 'bs')
plt.plot(seq, seq3, 'g^')
plt.show()

PRACTICE C

To plot points, we need one sequence of all the x values and one sequence of all the y values. Let's practice wrangling data to the format we need. The built-in zip function and splat syntax are our friends. Remember that using the splat syntax (*) with zip is the inverse of zip.

print(seq)
# [1, 2, 3, 4, 5, 6]
print(seq2)
# [1, 4, 9, 16, 25, 36]
zipped = list(zip(seq, seq2))
print(zipped)
# [(1, 1), (2, 4), (3, 9), (4, 16), (5, 25), (6, 36)]
unzipped = list(zip(*zipped))
print(unzipped)
# [(1, 2, 3, 4, 5, 6), (1, 4, 9, 16, 25, 36)]

Finish the following script by using zip on c.most_common() to plot the frequencies of the words in this quotation.

from collections import Counter
vote = 'Not voting is not a protest . It is a surrender .'.split()
print(vote)
# ['Not', 'voting', 'is', 'not', 'a', 'protest', '.', 'It', 'is', 'a', 'surrender', '.']
c = Counter(vote)  # dictionary
print(c)
# Counter({'is': 2, 'a': 2, '.': 2, 'Not': 1, 'voting': 1, 'not': 1, 'protest': 1, 'It': 1, 'surrender': 1})
print(c.most_common())
# [('is', 2), ('a', 2), ('.', 2), ('Not', 1), ('voting', 1), ('not', 1), ('protest', 1), ('It', 1), ('surrender', 1)]
...

You can use `plt.xticks(fontsize=10, rotation='vertical')

drawing barplot, scatterplot, and boxplots

import matplotlib.pyplot as plt

names = ['group_a', 'group_b', 'group_c']
values = [1, 10, 100]

plt.bar(names, values)
plt.show()

plt.scatter(names, values)
plt.show()

plt.plot(names, values)
plt.show()

plt.boxplot([1, 4, 4, 4, 5, 5, 5, 6, 6, 6, 7, 8, 11])
plt.show()

plt.boxplot([[1, 4, 4, 4, 5, 5, 5, 6, 6, 6, 7, 8, 11],
             [1, 1, 1, 1, 1, 2, 2, 2, 4, 7, 8, 9, 13]])
plt.show()

drawing a histogram

import matplotlib.pyplot as plt
import numpy as np

# Generate a normal distribution
mu, sigma = 100, 15
x = mu + sigma * np.random.randn(10000)  # x is a list of random numbers in a normal distribution

# the histogram of the data
counts, bins = np.histogram(x, bins=20)
plt.stairs(counts, bins, fill=True)  # or plt.bar()

# add some text
plt.xlabel('Smarts')
plt.ylabel('Probability')
plt.title('Histogram of IQ')
plt.text(50, 1200, r'$\mu=100,\ \sigma=15$')  # LaTeX math expression inside $'s
plt.grid(True)
plt.show()