Nested JSON in customised format from pandas Dataframe, with added label

Question

Dataframe

df = {"UNIT":["UNIT1","UNIT1","UNIT2","UNIT2"],
"PROJECT":["A","A","C","C"],
"TEAM":[1,2,1,2],
"NAME":["FANNY", "KATY", "PERCY", "PETER"],
"ID":[123,234,333,222]}
data = pd.DataFrame(df)

    UNIT PROJECT  TEAM   NAME   ID
0  UNIT1       A     1  FANNY  123
1  UNIT1       A     2   KATY  234
2  UNIT2       C     1  PERCY  333
3  UNIT2       C     2  PETER  222

Expected output

[
    {
        "UNIT": "UNIT1",
        "PROJECT": "A",
        "TEAM_DETAIL": [
            {
                "TEAM": 1,
                "MEMBER": [
                    {
                        "NAME": "FANNY",
                        "ID": 123
                    }
                ]
            },
            {
                "TEAM": "TEAM 2",
                "MEMBER": [
                    {
                        "NAME": "KATY",
                        "ID": 234
                    }
                ]
            }
        ]
    },
    {
        "UNIT": "UNIT2",
        "PROJECT": "C",
        "TEAM_DETAIL": [
            {
                "TEAM": 1,
                "MEMBER": [
                    {
                        "NAME": "PERCY",
                        "ID": 333
                    }
                ]
            },
            {
                "TEAM": "TEAM 2",
                "MEMBER": [
                    {
                        "NAME": "PETER",
                        "ID": 222
                    }
                ]
            }
        ]
    }
]

In this situation I would like to group the data by TEAM and hence showing each of the member details in each team. Without adding customised label eg.TEAM_DETAIL and MEMBER, it can be easily achieved by using .to_dict() However, I have no idea how to add a label on each level.