How do I transform df using pandas

Question

I have the following dataframe snippet:

predictor   b   z   pvalue  model   ss
age_raw 1.026695    4.678675    2.89E-06    composite_outcome_nonenglish    1
elixsum 1.228125    8.514571    1.67E-17    composite_outcome_nonenglish    1
age_raw 1.087716    0.228507    0.819252    composite_outcome_english   0
elixsum 1.760882    1.68492     0.092004    composite_outcome_english   0

that I need to transpose into a multi-index column dataframe using model as the highest level and (b, z, pvalue, ss) as the secondary level the predictor as rows:

                    model                              model
          composite_outcome_nonenglish            composite_outcome_english 
           b        z           pvalue   ss           b         z           pvalue  ss
age_raw 1.026695    4.678675    2.89E-06 0            1.087716  0.228507    0.819252 1
elixsum 1.228125    8.514571    1.67E-17 0            1.760882  1.68492     0.092004 1  

I've tried all sorts of groupings, and stacking and unstacking, etc., and for the life of me, I cannot get this right.

Answer 1

set index with unstack + stack

out = df.set_index(['predictor','model']).stack().unstack(level=[1,2])
Out[366]: 
model     composite_outcome_nonenglish            ... composite_outcome_english     
                                     b         z  ...                    pvalue   ss
predictor                                         ...                               
age_raw                       1.026695  4.678675  ...                  0.819252  0.0
elixsum                       1.228125  8.514571  ...                  0.092004  0.0