adding hatches to seaborn heatmap plot

问题

Is there a way to hatch particular 'cells' in a seaborn heatmap, which e.g. fullfill a condition? I already tried it with masked arrays and matplotlib pcolor, but it turned out that it hatched the wrong cells.

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

flights = sns.load_dataset("flights")
flights = flights.pivot("month", "year", "passengers")
zm = np.ma.masked_less(flights.values, 200)
x= np.arange(0,12)
y= np.arange(0,12)
sns.heatmap(flights,linewidth=.1)
plt.pcolor(x, y, zm, hatch='//', alpha=0.)

plt.show()

回答1:

I think the idea/strategy is correct. You just didn't use the correct coordinates.

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

flights = sns.load_dataset("flights")
flights = flights.pivot("month", "year", "passengers")
print(flights)
zm = np.ma.masked_less(flights.values, 200)
x= np.arange(len(flights.columns)+1)
y= np.arange(len(flights.index)+1)
sns.heatmap(flights,linewidth=.1)
plt.pcolor(x, y, zm, hatch='//', alpha=0.)

plt.show()

Potentially, one might want to hatch individual cells, instead of the complete area. This is hard. A wordaround is to create a grid of minor ticks and colorize it in white as to overlay the hatching.

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

plt.rcParams["axes.axisbelow"] = False

flights = sns.load_dataset("flights")
flights = flights.pivot("month", "year", "passengers")

zm = np.ma.masked_less(flights.values, 200)
x= np.arange(len(flights.columns)+1)
y= np.arange(len(flights.index)+1)

fig, ax = plt.subplots()
sns.heatmap(flights,linewidth=0, ax=ax)

ax.pcolor(x, y, zm, hatch='//', alpha=0.)

ax.set_xticks(np.arange(flights.shape[1]+1), minor=True)
ax.set_yticks(np.arange(flights.shape[1]+1), minor=True)
ax.grid(True, which="minor", color="w", linewidth=2)
ax.tick_params(which="minor", left=False, bottom=False)

plt.show()

来源：https://stackoverflow.com/questions/55285013/adding-hatches-to-seaborn-heatmap-plot