How to combine the values of Pandas dataframe rows by a shared attribute in another column? And produce a new column with this concatenated?

I am compiling a 'colleague award' list at my place of work. Members of staff can nominate as many colleagues as they like. What I would like is to produce a 'mail merge' style output where each staff member gets a list of their comments and who left them.

I have a .csv file which consists of 'Name' (of the person nominating), 'Who' (the person being nominated) and 'Why' (the comment).

How can I use PANDAS to combine the rows of the df by the common value, which is who the comment is about, and concatenate these together with who said it, into a new column in the DF with this info.

Code so far:

import numpy as np
import pandas as pd

df = pd.read_csv('Colleague Award 2023(1-296).csv')

df2 = df.drop(['ID', 'Start time' , 'Completion time', 'Email'], axis = 1)
df3 = df2.sort_values('Who', ascending=True)

df4 = df3.groupby(['Name'], as_index=False).agg({'Why' : ' '.join})

df4.to_csv('Colleage Award Collated.csv')

My output is just the 'why' strings stuck together and the 'who' column has been omitted.

EDIT/Update:
What I am really after is to write to a csv where the columns are like this:

But where all comments are in the same cell, not seperated

This is so I can then easily do a mail-merge and create many word documents, one for each recipient with all of their comments from various colleagues.

Thank you again to those who have already contributed!

Assuming this input format:

          Name          Who                               Why
0   Joe Bloggs     Jane Doe              Joe is super helpful
1   Joe Bloggs  Gary Public  Joe always brings my good coffee
2  Gary Public     Jane Doe   Gary always helps me with my PC
3  Gary Public   Joe Bloggs   Thank you Gary for training me!

You can use a loop with groupby:

sep = ''
for name, g in df.groupby('Name', sort=False):
    if sep:
        print(sep)
    sep = 'NEW ROW'
    print(f'{name} your comments:')
    for who, why in zip(g['Who'], g['Why']):
        print(f'{who}: {why}')

Output:

Joe Bloggs your comments:
Jane Doe: Joe is super helpful
Gary Public: Joe always brings my good coffee
NEW ROW
Gary Public your comments:
Jane Doe: Gary always helps me with my PC
Joe Bloggs: Thank you Gary for training me!

output as DataFrame

To create a new DataFrame:

out = df.assign(
          Why=lambda d: d['Who']+': '+d['Why'],
          Name=lambda d: d['Name'].mask(d['Name'].duplicated(), '')
         )[['Name', 'Why']]

To modify the input DataFrame in place:

df['Why'] = df.pop('Who')+': '+df['Why']
df.loc[df['Name'].duplicated(), 'Name'] = ''

Output:

          Name                                            Why
0   Joe Bloggs                 Jane Doe: Joe is super helpful
1               Gary Public: Joe always brings my good coffee
2  Gary Public      Jane Doe: Gary always helps me with my PC
3                 Joe Bloggs: Thank you Gary for training me!

To achieve the desired output, you can modify your code as follows:

import pandas as pd

df = pd.read_csv('Colleague Award 2023(1-296).csv')

df2 = df.drop(['ID', 'Start time', 'Completion time', 'Email'], axis=1)
df3 = df2.sort_values('Who', ascending=True)

df4 = df3.groupby(['Who'], as_index=False).agg({'Name': ' '.join, 'Why': lambda x: '\n'.join(x)})
df4['Comments'] = df4['Name'] + ': ' + df4['Why']
df4 = df4[['Who', 'Comments']]

df4.to_csv('Colleague Award Collated.csv', index=False)

The output file (Colleague Award Collated.csv) generated by the updated code will have two columns: 'Who' and 'Comments'. Each row represents a person being nominated, and the corresponding comments will be concatenated in the 'Comments' column, including the names of the nominators.