Seaborn heatmap xy ticks has unpropper date format

Question

I have big sparce pandas dataframe and using seaborn.heatmap to visualize it. The problem is I cannot have proper ticks on x and y axis - format is unreadable! On the x axis we have column names, which is strings and in the y axis it is index of datetime64 type. I tried code like

import matplotlib.dates as mdates

f, ax = plt.subplots(1, figsize=(8,8))
sns.heatmap(pvt.iloc[:,:], ax=ax)

ax.yaxis.set_major_formatter(mdates.DateFormatter("%Y-%m"))
ax.yaxis.set_minor_formatter(mdates.DateFormatter("%Y-%m"))

but this causes an error

Cannot convert 0 to a date. This often happens if non-datetime values are passed to an axis that expects datetime objects.

I think the problem is index has datetime64 type, but cannot handle with this. Please, help.

How it looks now:

seaborn.heatmap is a categorical plot. If you want a numeric plot, consider matplotlib imshow or pcolormesh. — ImportanceOfBeingErnest, Feb 13 '19 at 14:55
I'm happy with categorical plot. I just need to format dates properly. — user1443993, Feb 13 '19 at 15:28
Then format them in your dataframe. I.e. if the index is already strings in the format you want, they will not be changed by seaborn. See e.g [this](https://stackoverflow.com/questions/30132282/datetime-to-string-with-series-in-python-pandas) — ImportanceOfBeingErnest, Feb 13 '19 at 15:32
Actially, I have no idea how to do this. Y axis is an index, datetime64ns type, how to format it for heatmap? — user1443993, Feb 14 '19 at 14:01
See https://stackoverflow.com/questions/42185328/format-pandas-dataframe-index-date — ImportanceOfBeingErnest, Feb 14 '19 at 14:58

Seaborn heatmap xy ticks has unpropper date format

0 Answers0