I have a data frame which is described as below:
dataset method score input_shape
0 A dataset method1 0.9 K=64
0 A dataset method2 0.9 K=64
0 A dataset method1 0.9 K=32
0 A dataset method2 0.9 K=32
0 A dataset method3 0.9 L=1000
0 A dataset method3 0.9 L=2000
0 B dataset method1 0.9 K=64
0 B dataset method2 0.9 K=64
0 B dataset method1 0.9 K=32
0 B dataset method2 0.9 K=32
0 B dataset method3 0.9 L=1000
0 B dataset method3 0.9 L=2000
I am trying to use seaborn.catplot() to plot this datagram, to compare with performances under different dataset, and method, and input shape. Below is my code and the result I got:
sns.catplot(x='method', y='score', col="dataset", \
data=df, kind="bar",hue='input_shape', \
palette='Set2', height=5, aspect=1)
I observed that these bars are not centering with x-label due to some of my hue class is empty, but this result is not what I want. Is there any way to remove these empty class on the plot, making these bars can center with x-label (even without using catplot())?
Any suggestion is welcomed.
