Seaborn doesn't support PeriodIndex

Seaborn doesn’t properly support pandas PeriodIndex: dates should be converted to their actual values before being plotted.

import pandas as pd #Pandas 0.18.1
import seaborn as sns #Seaborn 0.7.0
index = pd.PeriodIndex(start='2000', end='2005', freq='A')
df = pd.DataFrame(np.random.randint(0,10,(len(index), 2)), index=index, columns=['A', 'B'])
sns.heatmap(df)

I’m not familiar with PRs, I guess there’s not much to change as converting the PeriodIndex to a regular datetime type does the trick sns.heatmap(df.pipe(lambda x: x.set_index(x.index.to_datetime())))

[davidgilbertson]

pandas: 1.4.2 matplotlib: 3.5.2 Python: 3.10.2

BTW Seaborn is great. Good docs too, things are nicely explained.

[mwaskom]

Weird, what version of pandas / matplotlib are you on?

Agreed that this is really something that should be made to work better in pandas.