Seaborn boxplots shifted incorrectly along x-axis

4
votes
from matplotlib import pyplot as plt
import pandas as pd
import seaborn as sns

df = pd.DataFrame({})
df[soi_name]=soi
df[outcome_name]=outcome
soi,outcome = utils.format_cols(soi, outcome,'continuous',agg_method)
sns.factorplot(data=df, x=outcome_name,y=soi_name,hue=outcome_name,kind='box')
plt.savefig(ofilepath)

enter image description here

So the code snippet used to generate this boxplot is just above. outcome is a binary float type pandas series. soi is a float type pandas series. This x-axis shift occurs for boxplots and violinplots. When I generate factor plots with the following code:

df = pd.DataFrame({})
df[soi_name]=soi
df[outcome_name]=outcome
sns.factorplot(data=df, x=outcome_name,y=soi_name,hue=outcome_name)
plt.savefig(ofilepath)

... enter image description here

I get my desired outputs. Any ideas with why the shift for the box plots might be happening?

1
pythonpandasmatplotlibseaborn

1 Answers

5
votes

You have two different hues, such that the first is plotted left of the position and the second is plotted right of the position. This is usually desired, see e.g. this plot from the documentation

enter image description here

import matplotlib.pyplot as plt
import seaborn as sns

tips = sns.load_dataset("tips")
ax = sns.boxplot(x="day", y="total_bill", hue="time",
                  data=tips, linewidth=2.5)

plt.show()

Here, you wouldn't want to have the blue and orange boxplots overlapping, would you?

But you can, if you want. Use the dodge=False argument for that purpose.

enter image description here

import matplotlib.pyplot as plt
import seaborn as sns

tips = sns.load_dataset("tips")
ax = sns.boxplot(x="day", y="total_bill", hue="time",
                  dodge=False, data=tips, linewidth=2.5)

plt.show()