Lesson 04

Multiple Lines and a Legend

Draw several lines on one set of axes, label each one, and place a legend so a reader can tell the lines apart without asking.

Intermediate25 min

Multiple Lines and a Legend hero Module 25 Lesson 4

Comparison is the whole point of most charts, and comparison needs more than one line. Call plot again and the second line lands on the same axes, in a new colour, sharing the same scale. The moment there are two lines, the reader needs a legend, and the legend comes from the label you give each line.

Three labelled lines sharing one set of axes with a legend

Call plot again

Each plot call adds a line to the current axes and picks the next colour in the default cycle. Nothing is replaced, so the axis range grows to fit whichever line goes highest.

python

import matplotlib.pyplot as plt

days = ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"]
python = [42, 58, 51, 66, 73, 61, 47]
numpy = [30, 35, 44, 41, 52, 48, 39]

plt.plot(days, python)
plt.plot(days, numpy)
print("lines on these axes:", len(plt.gca().lines))
plt.show()

Output:

text

lines on these axes: 2

Label each line, then ask for the legend

label= names a line, and plt.legend() collects every labelled line into a box. A line with no label is skipped, which is a neat way to keep a faint reference line out of the legend.

python

import matplotlib.pyplot as plt

days = ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"]
python = [42, 58, 51, 66, 73, 61, 47]
numpy = [30, 35, 44, 41, 52, 48, 39]
pandas_track = [18, 26, 31, 38, 47, 44, 36]

plt.plot(days, python, label="Python")
plt.plot(days, numpy, label="NumPy")
plt.plot(days, pandas_track, label="pandas")
plt.title("Signups by track")
plt.xlabel("day")
plt.ylabel("signups")
plt.legend()
plt.show()

for name, values in [("Python", python), ("NumPy", numpy), ("pandas", pandas_track)]:
    print(f"{name:<8} total {sum(values)}")

Output:

text

Python   total 398
NumPy    total 289
pandas   total 240

Where the legend sits

loc moves the legend. The default is "best", which hunts for the emptiest corner, and that is usually right. Name a corner yourself when you want the position to stay put across several charts, and give the box a title when the labels share a theme.

python

import matplotlib.pyplot as plt

days = ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"]
python = [42, 58, 51, 66, 73, 61, 47]
numpy = [30, 35, 44, 41, 52, 48, 39]

plt.plot(days, python, label="Python")
plt.plot(days, numpy, label="NumPy")
plt.legend(loc="upper left", title="Track")
plt.show()

print("some choices:", ", ".join(["best", "upper left", "lower right", "center"]))

Output:

text

some choices: best, upper left, lower right, center

Separate charts instead of one

show ends the current figure. Calling it between plots gives you two separate charts rather than two lines on one, which is what you want when the two measurements have nothing to do with each other or use wildly different scales.

python

import matplotlib.pyplot as plt

days = ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"]
signups = [42, 58, 51, 66, 73, 61, 47]
revenue = [1840, 2510, 2260, 2980, 3310, 2740, 2090]

plt.plot(days, signups)
plt.title("Signups")
plt.show()

plt.plot(days, revenue)
plt.title("Revenue")
plt.show()

print("plotting these together would flatten the signups line")

Output:

text

plotting these together would flatten the signups line

Both charts appear one after the other in the playground. Sharing one axis here would squash the signups line against the bottom, because revenue is fifty times larger.

Key takeaways

Each plot() call adds another line to the same axes
label= names a line and plt.legend() collects the labelled ones
loc moves the legend; the default 'best' finds an empty corner
Call show() between plots to get separate figures
Next: Line Styles and Colours
Previous: Your First Line Plot

Loading playground…

Check your understanding

Multiple Lines and a Legend Quiz

What happens on the second plot() call?

1 / 10

← Your First Line Plot

Line Styles and Colours →