Maybe this is a recurring problem here. Although I queried old questions on the site, I was unable to solve my problem completely. Some old forum topics helped me a lot:
Seaborn Heatmap with Datetime Axes
But after much trying I still could not figure it out.
My problem is this:
I have a dataframe like this (but the original is much more extensive, being the date index from 2004 to 2018 at one hour intervals).
Estreito Furnas Manso Itumbiara Curuai
2004-01-19 15:00:00 NaN NaN 3.73 NaN NaN
2004-01-19 16:00:00 NaN NaN 5.10 NaN NaN
2004-01-19 17:00:00 NaN NaN 5.10 NaN NaN
2004-01-19 18:00:00 NaN NaN NaN NaN NaN
2004-01-19 19:00:00 NaN NaN 4.31 NaN NaN
2004-01-19 20:00:00 NaN NaN 4.90 NaN NaN
2004-01-19 21:00:00 NaN NaN 5.88 NaN NaN
2004-01-19 22:00:00 NaN NaN 7.06 NaN NaN
2004-01-19 23:00:00 NaN NaN 4.71 NaN NaN
2004-01-20 00:00:00 NaN NaN NaN NaN NaN
2004-01-20 01:00:00 NaN NaN 7.25 NaN NaN
2004-01-20 02:00:00 NaN NaN 5.10 NaN NaN
2004-01-20 03:00:00 NaN NaN 2.55 NaN NaN
2004-01-20 04:00:00 NaN NaN 0.39 NaN NaN
2004-01-20 05:00:00 NaN NaN 0.20 NaN NaN
I'd like to show the missing data in a graph, so I'm using the heatmap of the seaborn module.
Following the documentation and old forum topics I get in the following script.
import matplotlib.pyplot as plt
import seaborn as sns
df_transposed = df.transpose()
plt.subplots(1,1,figsize=(12,6))
ax = sns.heatmap(df_transposed.isnull(), cbar=False)
x_dates = df.index.strftime('%Y-%m').sort_values().unique()
ax.set_xticklabels(labels=x_dates, rotation=45)
I got the following figure:
But as you can see, the xticklabels are not very good. And reading Seaborn's documentation I could not find any arguments to set the number of xticks.
I tried putting only years on xticklabels doing this:
x_dates = df.index.strftime('%Y').sort_values().unique()
But it doesn't work as well:
I humbly ask for help to solve this problem. I apologize if I am off topic.
Thank you very much in advance.
