CODEWITHSUNDEEP
pythonpandasheatmapseaborn
pho
pho

Reputation: 327

seaborn.heatmap skips data

When creating a heatmap via seaborn seaborn==0.7.0.dev0 my axis starts two hours later. The DataFrame used to create the heatmap starts at:

2015-05-19 21:10:00

The first get_xticklabels of the heatmap created via seaborn however is 2015-05-19 23:10:00.

The heatmap is created via

sns.heatmap(df_test.T, xticklabels=True, yticklabels=True, ax=ax)

What am I missing here? This example (using seaborn 0.7 and 0.6) will start one hour later, in my real data it is even 2 hours.

import seaborn as sns
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

index = pd.date_range('1/1/2000', periods=60*60*12, freq='s')


df = pd.DataFrame({'a': pd.Series(np.random.randn(len(index)), index=index),
                   'b': pd.Series(np.random.randn(len(index)), index=index)})

#create boolean vars
df.a = df.a > 0
df.b = df.b > 0

df = df.resample('1Min', how=np.mean)

ax = plt.gca()
fig = plt.gcf()


sns.heatmap(df.T, ax=ax)

#print index
print df.index[0]

#print first xlabel
print ax.get_xticklabels()[0]

[label.set_visible(False) for label in ax.xaxis.get_ticklabels()[1:]]
plt.gcf().autofmt_xdate()
plt.show()

This will result in the following output

user@debian:/tmp$ python test.py 
2000-01-01 00:00:00
Text(0.5,0,u'2000-01-01T01:00:00.000000000+0100')

Upvotes: 0

Views: 1274

Answers (1)

pho
pho

Reputation: 327

Okay turns out it is the missing tz= value when creating the index which gives the offset inb the example code.

My solution (as changing tz in my DataFrame did not change this behaviour) was to set xticklabel=False in heatmap() and use plt.xticks() directly.

Upvotes: 1

Related Questions