Bar chart - can we simplify?

📊 Plotly Python
1

I have a dataframe:

>>> import pandas as pd
>>> import plotly.express as px
>>> df = pd.DataFrame({
...     'P': ['P1', 'P2', 'P3'],
...     'Q1': [1, 2, 3],
...     'Q2': [3, 1, 2],
... })
>>> df
    P  Q1  Q2
0  P1   1   3
1  P2   2   1
2  P3   3   2
>>>

For each row, I should sort Qx columns in descending order, and than do plots but each Qx column should have the same color. Something like this:

qcols = [col for col in df.columns if col.startswith('Q')]

# I have to define colors in advance
color_dict = {col: color for col, color in zip(
    qcols,
    px.colors.qualitative.Plotly[:len(df)]
)}


for p in df.P.unique():
    # Transformation needed for sorting columns descending
    dfp = df.query("P == @p").set_index('P').T.sort_values(p, ascending=False).T.reset_index().melt(id_vars='P', var_name='Q', value_name='val')
    fig = px.bar(data_frame=dfp, x='Q', y='val', color='Q', color_discrete_map=color_dict, width=300, height=300)
    fig.show()

image
image490×1613 23.1 KB

It looks quite complicated to me.
Can it be done in a simpler way?