Reverse value (categorical) order of colorbar-CodePudding

I have a dataframe where energylabel is a categorical variable.

>>> df
    costs  EP2
0     0.0  260
1  5866.0  230
2  5755.0   32
3   722.0  -42
4  6588.0  103
>>> bins_energylabel = [-np.inf, 0, 50, 75, 105, 160, 190, 250, 290, 335, 380, np.inf]
>>> labs_energylabel = ['A    ', 'A   ', 'A  ', 'A ', 'A', 'B', 'C', 'D', 'E', 'F', 'G']
>>> df['energylabel'] = pd.cut(df['EP2'], bins=bins_energylabel, labels=labs_energylabel)
>>> df
    costs  EP2 energylabel
0     0.0  260           D
1  5866.0  230           C
2  5755.0   32        A   
3   722.0  -42       A    
4  6588.0  103          A

I want to plot the costs versus EP2 and coloring the energylabels with their corresponding energylabel.

ax1 = df.plot.scatter(x='costs', y='EP2', c='energylabel', colormap='Set3', sharex=False)

Looks something like this:

Now I want to reverse the order of the colormap values: instead of G on top, I want A etc.. My original dataframe is longer and contains more labels, the categories are ordered like so (this is done by using pd.cut() on column EP2).

Categories (11, object): ['A    ' < 'A   ' < 'A  ' < 'A ' ... 'D' < 'E' < 'F' < 'G']

I think by reversing the order it can solve my problem. But is it also possible to do this directly from the plotting code above?

CodePudding user response：

IIUC, you can reverse the colormap by adding an _r suffix and invert the colorbar Y-axis with

NB. inverting the colors is optional, if you don't, you will get: