Seaborn countplot set legend for x values

Question

I'm ploting a categorical data and value count by sns.countplot()

I'm trying to add legend for x-values to the figure as following: handles is set of x-value, labels is the descriptions of x-values.

ax = sns.countplot(x = df.GARAGE_DOM)
handles, labels = ax.get_legend_handles_labels()

handles = ["VP", "BC", "GC", "GP", "JC", "PO"]
labels = ["Voie Publique", "box", "Garage couvert", "garage particulier clos", "Jardin clos", "parking ouvert"]
by_label = OrderedDict(zip(handles,labels))
ax.legend(by_label.keys(), by_label.values())

However, I got warning that

UserWarning:

Legend does not support 'VP' instances. A proxy artist may be used instead. See: http://matplotlib.org/users/legend_guide.html#using-proxy-artist

I've read the doc of proxy artist but I didn't find examples in my case.

Thanks for your help.

Answer 1

Here is a possible solution, creating a text field as a legend handler. The following would create a TextHandler to be used to create the legend artist, which is a simple matplotlib.text.Text instance. The handles for the legend are given as tuples of (text, color) from which the TextHandler creates the desired Text.

import seaborn as sns
import matplotlib.pyplot as plt
from matplotlib.legend_handler import HandlerBase
from matplotlib.text import Text
import numpy as np
import pandas as pd

class TextHandler(HandlerBase):
    def create_artists(self, legend, tup ,xdescent, ydescent,
                        width, height, fontsize,trans):
        tx = Text(width/2.,height/2,tup[0], fontsize=fontsize,
                  ha="center", va="center", color=tup[1], fontweight="bold")
        return [tx]


a = np.random.choice(["VP", "BC", "GC", "GP", "JC", "PO"], size=100, 
                     p=np.arange(1,7)/21. )
df = pd.DataFrame(a, columns=["GARAGE_DOM"])

ax = sns.countplot(x = df.GARAGE_DOM)


handltext = ["VP", "BC", "GC", "GP", "JC", "PO"]
labels = ["Voie Publique", "box", "Garage couvert", "garage particulier clos", "Jardin clos", "parking ouvert"]


t = ax.get_xticklabels()
labeldic = dict(zip(handltext, labels))
labels = [labeldic[h.get_text()]  for h in t]
handles = [(h.get_text(),c.get_fc()) for h,c in zip(t,ax.patches)]

ax.legend(handles, labels, handler_map={tuple : TextHandler()}) 

plt.show()

---

The above solution is an updated version of the original version below, which seems more complicated. The following is the original solution, which uses a TextArea and an AnchoredOffsetbox to place the text inside the legend.

import seaborn.apionly as sns
import matplotlib.pyplot as plt
import matplotlib.patches as patches
from matplotlib.offsetbox import TextArea, AnchoredOffsetbox
from matplotlib.transforms import TransformedBbox, Bbox
from matplotlib.legend_handler import HandlerBase
import numpy as np
import pandas as pd

class TextHandler(HandlerBase):
    def __init__(self, text, color="k"):
        self.text = text 
        self.color = color
        super(TextHandler, self).__init__()

    def create_artists(self, legend, orig_handle,xdescent, ydescent,
                        width, height, fontsize,trans):
        bb = Bbox.from_bounds(xdescent,ydescent, width,height)
        tbb = TransformedBbox(bb, trans)
        textbox = TextArea(self.text, textprops={"weight":"bold","color":self.color})
        ab = AnchoredOffsetbox(loc=10,child=textbox, bbox_to_anchor=tbb, frameon=False)
        return [ab]


a = np.random.choice(["VP", "BC", "GC", "GP", "JC", "PO"], size=100, 
                     p=np.arange(1,7)/21. )
df = pd.DataFrame(a, columns=["GARAGE_DOM"])

ax = sns.countplot(x = df.GARAGE_DOM)


handltext = ["VP", "BC", "GC", "GP", "JC", "PO"]
labels = ["Voie Publique", "box", "Garage couvert", "garage particulier clos", "Jardin clos", "parking ouvert"]

handles = [ patches.Rectangle((0,0),1,1) for h in handltext]
t = ax.get_xticklabels()
labeldic = dict(zip(handltext, labels))
labels = [labeldic[h.get_text()]  for h in t]
handlers = [TextHandler(h.get_text(),c.get_fc()) for h,c in zip(t,ax.patches)]
handlermap = dict(zip(handles, handlers))
ax.legend(handles, labels, handler_map=handlermap,) 

plt.show()

---

Also see this more generic answer