How to create a nested dictionary from pandas DataFrame

Question

I'm not very practical with python data series, can someone explain what I'm doing wrong?

I have this DataFrame:

Model	Year	Brand	Number
Sorento	2008	Kia	19UYA427X1A382858
Corvette	1968	Chevrolet	1C3CDFEB2FD382486
Sienna	1933	Toyota	YV440MBK0F1112644
Corvette	1968	Kia	45UYA427X1A382858

and I need this output:

{
   "Kia": {
        "2008": [
            [
                "Sorento",
                "19UYA427X1A382858"
            ],
            [ 
                "Sorento",
                "45UYA427X1A382858"
            ]
        ]
    }, 
    "Chevrolet": {
      "1968": [ 
         [
          "Corvette",
          "1C3CDFEB2FD382486"
         ] 
       ]
    }
}

So I need to group the items with the same value into subList, but I'm not figuring out how to do it. The Model and Year are grouping correctly but I don't know how to fill the last list with the Brand and Number values

 ## this is my partial solution
d = {k: f.groupby('Year').apply(list).to_dict()
     for k, f in df_clean.groupby('Brand')}

pprint.pprint(d)

Output:
 'Mitsubishi': {1994.0: ['Model', 'Year', 'Brand', 'Number'],
                2005.0: ['Model', 'Year', 'Brand', 'Number'],
                'NO_DATA': ['Model', 'Year', 'Brand', 'Number']},
 'Nissan': {1993.0: ['Model', 'Year', 'Brand', 'Number'],
            1996.0: ['Model', 'Year', 'Brand', 'Number'],
            2009.0: ['Model', 'Year', 'Brand', 'Number'],
            2011.0: ['Model', 'Year', 'Brand', 'Number'],
            2012.0: ['Model', 'Year', 'Brand', 'Number']},
 'Oldsmobile': {1999.0: ['Model', 'Year', 'Brand', 'Number']},

Answer 1

You could use a nested groupby.apply; once for the brands and again for the years:

out = (df.groupby('Brand')
       .apply(lambda x: x.groupby('Year')[['Model','Number']]
              .apply(lambda y: y.to_numpy().tolist())
              .to_dict())
       .to_dict())

Output:

{'Chevrolet': {1968: [['Corvette', '1C3CDFEB2FD382486']]},
 'Kia': {1968: [['Corvette', '45UYA427X1A382858']],
  2008: [['Sorento', '19UYA427X1A382858']]},
 'Toyota': {1933: [['Sienna', 'YV440MBK0F1112644']]}}