12

Does anyone know if we can plot filled boxplots in python matplotlib? I've checked http://matplotlib.org/api/pyplot_api.html but I couldn't find useful information about that.

MHardy
  • 491
  • 3
  • 7
  • 17

2 Answers2

41

The example that @Fenikso shows an example of doing this, but it actually does it in a sub-optimal way.

Basically, you want to pass patch_artist=True to boxplot.

As a quick example:

import matplotlib.pyplot as plt
import numpy as np

data = [np.random.normal(0, std, 1000) for std in range(1, 6)]
plt.boxplot(data, notch=True, patch_artist=True)

plt.show()

enter image description here

If you'd like to control the color, do something similar to this:

import matplotlib.pyplot as plt
import numpy as np

data = [np.random.normal(0, std, 1000) for std in range(1, 6)]

box = plt.boxplot(data, notch=True, patch_artist=True)

colors = ['cyan', 'lightblue', 'lightgreen', 'tan', 'pink']
for patch, color in zip(box['boxes'], colors):
    patch.set_facecolor(color)

plt.show()

enter image description here

Joe Kington
  • 275,208
  • 71
  • 604
  • 463
  • Add `patch.set_alpha(0.5)` to the loop for transparency. – n1000 Jan 26 '15 at 15:47
  • 1
    @n1000 - Note that you can combine calls to "set_" by using the `set` method. For example: `patch.set(facecolor=color, alpha=0.5)`. – Joe Kington Jan 26 '15 at 19:28
  • This is great - thank you! I had searched for a clean way to do this for hours. – Arne Babenhauserheide May 29 '15 at 09:35
  • 2
    Is it possible to set transparency to the facecolor but not to the line of the box ? – Ger Oct 18 '15 at 22:53
  • @Ger: For that you'd probably have to do the "same" plot twice: Once with the patch artist, transparent fill, and no lines at all, and then a second time with the line artist (`patch_artist=False`, the default) to get the non-transparent lines. – Lutz Prechelt Jan 05 '23 at 10:26
2

You can do this with the Plotly Python API. The graph, script, and data for this graph are here.

To control color, you'll want to stipulate a fillcolor. Here, it's not set; the default is to fill it. Or, you can make it transparent, by adding 'fillcolor':'rgba(255, 255, 255, 0)'. You could also style with the GUI to tweak it.

import plotly
py = plotly.plotly(username='username', key='api_key')

from numpy.random import lognormal

x=[0]*1000+[1]*1000+[2]*1000
y=lognormal(0,1,1000).tolist()+lognormal(0,2,1000).tolist()+lognormal(0,3,1000).tolist()
s={'type':'box','jitter':0.5}
l={'title': 'Fun with the Lognormal distribution','yaxis':{'type':'log'}}

py.plot(x,y,style=s,layout=l)

Full disclosure: I'm on the Plotly team.

Boxplot

Mateo Sanchez
  • 1,604
  • 15
  • 20