A plotting library

Adding Text in Plots

You already know how to add legends to charts, but what about other text, maybe in nonstandard parts of a plot? Listing 8 shows the simplest way to add the captions and annotations shown in Figure 6.

Listing 8

Inserting Text

01 import Matplotlib.pyplot as plt
02 import NumPy as np
04 t = np.arange(0., 50, 0.2)
07 plt.plot(t, (t**3)*np.cos(2*t))
09 plt.title("Combination of trigonometric and power factors")
11 plt.text(1, 80000, r'$Function: (t^\mu)*cos(\sigma*t),\ with\ \mu = 3\ and\  \sigma = 2$')
13 plt.annotate('local maximum', xy=(19, 4100), xytext=(1, 45000), arrowprops=dict(facecolor='red', shrink=0.05))
15 plt.savefig('text-in-plots.png')
Figure 6: If labels and legends aren't enough, you also can add captions and annotations where you want.

Lines 11 and 13 of Listing 8 describe the function that generates the "local maximum" annotation shown in Figure 6. This syntax is relatively simple to describe: plt.annotate takes as arguments the string to insert, its starting position (xytext), the point where the arrow should end (xy), and its style (arrowprops). In line 11, plt.text has even fewer arguments: just the coordinates where a string should be inserted and then the string itself. The string is enclosed by single quotes and dollar signs and preceded by an r suffix, which means that what follows is "raw" text that uses LaTeX syntax and symbols (e.g., \mu and \sigma). Also note that in these strings blank spaces must be escaped to tell Python that they are actual spaces that should just be printed instead of separators of function arguments.

Categorical Charts

Another place in a plot where you might want arbitrary text instead of numbers or dates is on its Axes. In Matplotlib, this type of diagram is called a categorical chart. Figure 7 shows a categorical chart of how cats and dogs consider certain activities.

Figure 7: If they are properly categorized, charts can connect and display data of all kinds.

Unless you have lots of categories to display or other special needs, plotting categorical charts is easier than it appears. Listing 9 only needs three arrays of strings and five lines of code to create Figure 7. The activity array lists all the activities, and the cat and dog arrays describe how those pets consider each activity (lines 3 to 5). The only critical part here is that each reaction of cats or dogs must be in the same position of the corresponding activity. That is, if playing is the fifth element of the activity array, and dogs are SUPER-HAPPY only when playing, then SUPER-HAPPY must be the fifth element of the dog array and the only element with the SUPER-HAPPY value. Then, all Matplotlib needs to plot everything correctly are the ax.plot methods shown in lines 8 and 9, which take the two arrays as the lists of x and y values to use to find all the points of each plot.

Listing 9

Categorical Chart

01 import Matplotlib.pyplot as plt
03 activity = ["combing", "drinking", "feeding", "napping", "playing", "washing"]
04 cat = ["bored", "happy", "SUPER-HAPPY", "SUPER-HAPPY", "happy", "bored"]
05 dog = ["bored", "happy", "happy", "bored", "SUPER-HAPPY", "bored"]
07 fig, ax = plt.subplots()
08 ax.plot(activity, dog, label="dog")
09 ax.plot(activity, cat, label="cat")
10 ax.legend()
11 plt.title("Pet feelings vs pet activities")
12 plt.savefig('categorical-names.png')

Sankey Diagrams

A Sankey diagram [7] represents the flow of certain variables in or out of a system by depicting each flow's width proportional to its quantity. While this type of diagram is not useful for everyone, a Sankey diagram showcases Matplotlib's flexibility and might also be a lifesaver for those in the fields of energy management, manufacturing, and science. Figure 8 and its code in Listing 10 are another example provided, but not completely explained, by the official Matplotlib documentation [8].

Listing 10

Sankey Diagram

01 import Matplotlib.pyplot as plt
02 from Matplotlib.sankey import Sankey
04 Sankey(flows=[0.25, 0.15, 0.60, -0.20, -0.15, -0.05, -0.50, -0.10],
        labels=['', '', '', 'First', 'Second', 'Third', 'Fourth', 'Fifth'], orientations=[-1, 1, 0, 1, 1, 1, 0, -1]).finish()
05 plt.title("Did you know how easy it is to create diagrams like this?")
06 plt.savefig('sankey.png')
Figure 8: Sankey diagrams visualize input and output flows of materials, energy, or other quantities.

To draw a Sankey diagram, Matplotlib (or any other tool, for that matter) needs two things: a function, or object, that knows how to draw a Sankey diagram, and the parameters of each flow that enters or leaves the system. Listing 10 does just that by importing the Sankey function (line 2) and then telling Matplotlib to create and finish a Sankey object with the single long command in line 4. To do that, of course, each flow must be described by three parameters, namely the flow's intensity, orientation (i.e., if it enters or leaves the system), and an optional label. Lines 4 and 6 of Listing 10 pass this data passed to the Sankey object. As an example, the arrow pointing downward in Figure 8 (with a value of 0.1 and the label "Fifth") is the direct result of writing -0.10, 'Fifth', and -1 in the last elements of the flows, labels, and orientation arrays (lines 4-6) passed to the Sankey object.

Buy this article as PDF

Express-Checkout as PDF
Price $2.95
(incl. VAT)

Buy Linux Magazine

Get it on Google Play

US / Canada

Get it on Google Play

UK / Australia

Related content

  • Data Visualization in Python

    Python's powerful Matplotlib, Bokeh, PyQtGraph, and Pandas libraries lend programmers a helping hand when visualizing complex data and their relationships.

  • Unsupervised Learning

    The most tedious part of supervised machine learning is providing sufficient supervision. However, if the samples come from a restricted sample space, unsupervised learning might be fine for the task.

  • PyScript

    PyScript lets you use your favorite Python libraries on client-side web pages.

  • Tutorial – Prettymaps

    Prettymaps combines multiple Python libraries to make it easy to draw maps straight from the OpenStreetMap database.

  • Machine Learning

    We explore some machine learning techniques with a simple missing person app.

comments powered by Disqus
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Subscribe to our ADMIN Newsletters

Support Our Work

Linux Magazine content is made possible with support from readers like you. Please consider contributing when you’ve found an article to be beneficial.

Learn More