Seaborn heatmap is generating additional ticks on colorbar when using log scale

问题

I am trying to make a heatmap with logarithmic colorbar. But it keeps generating its own ticks and ticklabels along with the ones I input.

I originally posted this to reformat the tick labels from scientific notation to plain but then ran into this problem.

import numpy as np
import seaborn as sns
from matplotlib.colors import LogNorm
import matplotlib.ticker as tkr

matrix = np.random.rand(10, 10)/0.4
vmax=2
vmin=0.5

cbar_ticks = [0.5, 0.75, 1, 1.33, 2]
formatter = tkr.ScalarFormatter(useMathText=True)
formatter.set_scientific(False)

log_norm = LogNorm(vmin=vmin, vmax=vmax)
ax = sns.heatmap(matrix, square=True, vmax=vmax, vmin=vmin, norm=log_norm, cbar_kws={"ticks": cbar_ticks, "format": formatter})

回答1:

With a logarithmic axis, often also minor ticks are set (they help to know where different values are situated and they enforce the logarithmic look). In this case, the default ticks only include one tick (at 1.0) which isn't enough to see which value correspond to which color.

With cbar_kws only the major ticks can be changed. You can suppress the minor ticks explicitly:

import numpy as np
import seaborn as sns
from matplotlib.colors import LogNorm
import matplotlib.ticker as tkr
from matplotlib import pyplot as plt

matrix = np.random.rand(10, 10) / 0.4
vmax = 2
vmin = 0.5

cbar_ticks = [0.5, 0.75, 1, 1.33, 2]
formatter = tkr.ScalarFormatter(useMathText=True)
formatter.set_scientific(False)

log_norm = LogNorm(vmin=vmin, vmax=vmax)
ax = sns.heatmap(matrix, square=True, vmax=vmax, vmin=vmin, norm=log_norm,
                 cbar_kws={"ticks": cbar_ticks, "format": formatter})
ax.collections[0].colorbar.ax.yaxis.set_ticks([], minor=True)
plt.show()

来源：https://stackoverflow.com/questions/65675891/seaborn-heatmap-is-generating-additional-ticks-on-colorbar-when-using-log-scale

标签

python

matplotlib

seaborn