Applying "pivot_table" to df, I obtain a mean.
df_mean=df.pivot_table(index=[\'col_1\']).sort_index(axis=1, level=1)
when I show the WI
