seaborn barplot with labels for x values (and no hue)

Question

My dataframe contains two columns, I would like to plot their values in a barplot. Like this:

import seaborn as sns

# load sample data and drop all but two columns
tips = sns.load_dataset("tips")
tips= tips[["day", "total_bill"]]

sns.set(style="whitegrid")
ax = sns.barplot(x="day", y="total_bill", data=tips)

On top of this barplot, I would also like to add a legend with labels for each x value. Seaborn supports this, but as far as I can see, it works only when you specify a hue argument. Each label in the legend then corresponds to a hue value.

Can I create a legend with explanations for the x values? This might be a confusing question. I don't want to rename the label for the axis or the ticks along the axis. Instead, I would like to have a separate legend with additional explanations. My bars give me some nice space to put this legend and the explanations would be too long to have them as ticks.

Quang Hoang Quang Hoang · Accepted Answer · 2019-09-30T15:20:26

Is this what you want:

sns.set(style="whitegrid")
ax = sns.barplot(x="day", y="total_bill", data=tips)

ax.legend(ax.patches, ['1','2','3','Something that I can\'t say'], loc=[1.01,0.5])

Output:

seaborn barplot with labels for x values (and no hue)

1 Answers