How to alter the hue legend? - python-3.x

Question

I have a dataset that represents male and female in binary. Males are represented as 0 while females are represented as 1. What i hope to do is to change 0 to Male, and 1 to Female in the plot legend. I tried to follow this post, but it didn't work out.
It gives me an error message that looks like this:
AttributeError Traceback (most recent call last)
<ipython-input-11-b3c99d4311ab> in <module>
23 # plot the legend
24 plt.legend()
---> 25 legend = g._legend
26 new_labels = ['Female', 'Male']
27 for t, l in zip(legend.texts, new_labels): t.set_text(l)
AttributeError: 'AxesSubplot' object has no attribute '_legend'
This is how my currrent code looks like:
## store them in different variable names
X = salary['years']
y = salary['salary']
g = salary['gender']
# prepare the scatterplot
sns.set()
plt.figure(figsize=(10,10))
g = sns.scatterplot(x=salary.years, y=salary.salary, data=salary, hue='gender')
# equations of the models
model1 = 50 + 2.776962335386217*X
model2 = 60.019802 + 2.214645*X
model3_male = 60.014922 + 2.179305*X + 1.040140*1
model3_female = 60.014922 + 2.179305*X + 1.040140*0
# plot the scatterplots
plt.plot(X, model1, color='r', label='Model 1')
plt.plot(X, model2, color='g', label='Model 2')
plt.plot(X, model3_male, color='b', label='Model 3(Male)')
plt.plot(X, model3_female, color='y', label='Model 3(Female)')
# plot the legend
plt.legend()
legend = g._legend
new_labels = ['Female', 'Male']
for t, l in zip(legend.texts, new_labels): t.set_text(l)
# set the title
plt.title('Scatterplot of salary and model fits')
plt.show()

Answer 1

I don't have your data, so I generate some by my own:
gender salary years
male 40000 1
male 32000 2
male 45000 3
male 54000 4
female 72000 5
female 62000 6
female 92000 7
female 55000 8
female 35000 9
female 48000 10
import seaborn as sns
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
salary = pd.read_csv("1.csv", delim_whitespace=True)
print(salary)
X = salary['years']
y = salary['salary']
g = salary['gender']
# prepare the scatterplot
sns.set()
plt.figure(figsize=(10,10))
g = sns.scatterplot(x=salary.years, y=salary.salary, data=salary, hue='gender')
# equations of the models
model1 = 50 + 2.776962335386217*X
model2 = 60.019802 + 2.214645*X
model3_male = 60.014922 + 2.179305*X + 1.040140*1
model3_female = 60.014922 + 2.179305*X + 1.040140*0
# plot the scatterplots
plt.plot(X, model1, color='r', label='Model 1')
plt.plot(X, model2, color='g', label='Model 2')
plt.plot(X, model3_male, color='b', label='Model 3(Male)')
plt.plot(X, model3_female, color='y', label='Model 3(Female)')
# plot the legend
plt.legend()
# set the title
plt.title('Scatterplot of salary and model fits')
plt.show()
It works fine. So I guess values in your gender column are 0 or 1. In that case, you can do the following before g = salary['gender'] to replace 0 with male and 1 with female:
salary['gender'] = salary['gender'].map({1: 'female', 0: 'male'})
Back to your error:
---> 25 legend = g._legend
26 new_labels = ['Female', 'Male']
27 for t, l in zip(legend.texts, new_labels): t.set_text(l)
AttributeError: 'AxesSubplot' object has no attribute '_legend'
g returned by sns.scatterplot is class matplotlib.axes.Axes. To get lengend object from it, you need to use ax.get_legend() or ax.legend() rather than ax._legend. You can follow the officail Legend guide documentation.
legend = g.legend()
new_labels = ['Female', 'Male']
for t, l in zip(legend.texts[-2:], new_labels): t.set_text(l)