Increment matplotlib color cycle

Question

Is there a simple way to increment the matplotlib color cycle without digging into axes internals?

When plotting interactively a common pattern I use is:

import matplotlib.pyplot as plt

plt.figure()
plt.plot(x,y1)
plt.twinx()
plt.plot(x,y2)

The plt.twinx() in necessary to get different y-scales for y1 and y2 but both plots are drawn with the first color in the default colorcycle making it necessary to manually declare the color for each plot.

There must be a shorthand way to instruct the second plot to increment the color cycle rather than explicitly giving the color. It is easy of course to set color='b' or color='r' for the two plots but when using a custom style like ggplot you would need need to lookup the color codes from the current colorcycle which is cumbersome for interactive use.

unutbu · Accepted Answer · 2016-06-18 10:02:12Z

16

You could call

ax2._get_lines.get_next_color()

to advance the color cycler on color. Unfortunately, this accesses the private attribute ._get_lines, so this is not part of the official public API and not guaranteed to work in future versions of matplotlib.

A safer but less direct way of advance the color cycler would be to plot a null plot:

ax2.plot([], [])

import numpy as np
import matplotlib.pyplot as plt

x = np.arange(10)
y1 = np.random.randint(10, size=10)
y2 = np.random.randint(10, size=10)*100
fig, ax = plt.subplots()
ax.plot(x, y1, label='first')
ax2 = ax.twinx()
ax2._get_lines.get_next_color()
# ax2.plot([], [])
ax2.plot(x,y2, label='second')

handles1, labels1 = ax.get_legend_handles_labels()
handles2, labels2 = ax2.get_legend_handles_labels()
ax.legend(handles1+handles2, labels1+labels2, loc='best')  

plt.show()

edited Jun 18, 2016 at 10:02

answered Jun 17, 2016 at 22:15

unutbu

886k197 gold badges1.9k silver badges1.7k bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Mike Over a year ago

This solves the problem, but this is just the kind of verbosity that I was trying to avoid. Explicitly using the axes object to configure the plot is probably a good idea in general but I still find myself using the plt.plot() shortcuts that write to the gca() quite often.

Mike Over a year ago

Actually, I misread the answer. You do seem to change the color in one line. The complicated part is getting the legend to work for the twin axes.

Mike Over a year ago

It would be nice to have a factory like plt.subplots() that can generate several axes in the same canvas with things like colors and legends cooperating intuitively.

Mike Over a year ago

I did not see an ax._get_lines._get_next_color() method in matplotlib 1.5.1. The null plot solution is very simple though and does what I need.

Michele · Accepted Answer · 2019-07-26 14:41:21Z

15

Similar to the other answers but using matplotlib color cycler:

import matplotlib.pyplot as plt
from itertools import cycle

prop_cycle = plt.rcParams['axes.prop_cycle']
colors = cycle(prop_cycle.by_key()['color'])
for data in my_data:
    ax.plot(data.x, data.y, color=next(colors))

answered Jul 26, 2019 at 14:41

Michele

3,1412 gold badges25 silver badges34 bronze badges

1 Comment

bluenote10 Over a year ago

+1 because side effect free, avoids private _get_lines access, and makes it quite clear where the color values are read from.

Joe Heffer · Accepted Answer · 2020-04-01 15:01:48Z

7

There are several colour schemes available in Pyplot. You can read more on the matplotlib tutorial Specifying Colors.

From these docs:

a "CN" color spec, i.e. 'C' followed by a number, which is an index into the
default property cycle (matplotlib.rcParams['axes.prop_cycle']); the indexing
is intended to occur at rendering time, and defaults to black if the cycle
does not include color.

You can cycle through the colour scheme as follows:

fig, ax = plt.subplots()

# Import Python cycling library
from itertools import cycle

# Create a colour code cycler e.g. 'C0', 'C1', etc.
colour_codes = map('C{}'.format, cycle(range(10)))

# Iterate over series, cycling coloour codes
for y in my_data:
    ax.plot(x, y, color=next(color_codes))

This could be improved by cycling over matplotlib.rcParams['axes.prop_cycle'] directly.

edited Apr 1, 2020 at 15:01

answered Aug 24, 2018 at 13:14

Joe Heffer

8458 silver badges10 bronze badges

2 Comments

Bananach Over a year ago

+1 For not relying on side effects. Maybe for completeness your snippet could end with ax.plot(x,y,color=next(color_codes)) instead

bluenote10 Over a year ago

Where are the colors actually coming from? Python 2 user warning: map('C{}'.format, cycle(range(10))) is an infinite loop.

Collectives™ on Stack Overflow

Increment matplotlib color cycle

3 Answers 3

4 Comments

1 Comment

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

4 Comments

1 Comment

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related