How to to fix legend colors with x_bins in seaborn regplot?

Question

Seaborn regplot stops matching legend color to line color when I include x_bins in the parameters. It works fine until I add x_bins and then the multi-colored legend loses its color differentiation.

import matplotlib.pyplot as plt
import numpy as np; np.random.seed(1)
import pandas as pd
import seaborn as sns

data=pd.DataFrame({"VarX" : np.arange(10), 
                   'VarY1': np.random.rand(10),
                   'VarY2': np.random.rand(10),
                   'VarY3': np.random.rand(10)})

fig = plt.figure(figsize=(10,6))
sns.regplot(x='VarX', y='VarY1', data=data, x_bins=10)
sns.regplot(x='VarX', y='VarY2', data=data, x_bins=10)
sns.regplot(x='VarX', y='VarY3', data=data, x_bins=10)
fig.legend(labels=['First','Second','Third'])
plt.show()

Answer 1

Seaborn has its own concept of legend which often conflicts with default matplotlib legends.

To stay in the seaborn way of thinking, you can use lmplot for this and let it automatically produce the legend. This requires to reshape the input data a bit.

import seaborn as sns
import matplotlib.pyplot as plt
import numpy as np; np.random.seed(1)
import pandas as pd


data=pd.DataFrame({"VarX" : np.arange(10), 
                   'VarY1': np.random.rand(10),
                   'VarY2': np.random.rand(10),
                   'VarY3': np.random.rand(10)})

df = data.set_index("VarX")
df.columns = ['First','Second','Third']
df = df.stack().rename_axis(['VarX','Ycategory']).rename('VarY').reset_index()


sns.lmplot(x="VarX", y="VarY", hue="Ycategory", data = df, x_bins=10)

plt.show()