Seaborn 线图意外行为

Question

I am hoping to understand why the following Seaborn lineplot behaviour occurs.

Spikes are occurring through the time-series and additional data has been added to the left of the actual data.

How can I prevent this unexpected behaviour in Seaborn?

Regular plot of data:

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

aussie_property[['Sydney(SYDD)']].plot();

Seaborn plot of data:

sns.lineplot(data=aussie_property, x='date', y='Sydney(SYDD)');

Answer 1

This is not a seaborn problem but a question of ambiguous datetimes.

Convert date to a datetime object with the following code:

aussie_property['date'] = pd.to_datetime(aussie_property['Date'], dayfirst=True)

and you get your expected plot with seaborn

Generally, it is advisable to provide the format during datetime conversions, eg,

aussie_property['date'] = pd.to_datetime(aussie_property['Date'], format="%d/%m/%Y")

because, as we have seen here, dates like 10/12/2020 are ambiguous. Consequently, the parser first thought the data would be month/day/year and later noticed this cannot be the case, so changed to parsing your input as day/month/year, giving rise to these time-travelling spikes in your seaborn graph. Why you didn't see them in the pandas plot, you ask? Well, this is plotted against the index, so you don't notice this conversion problem in the pandas plot.
More information on the format codes can be found in the Python documentation .