Python Percentage Change with group

I have a dataframe. And, I also have start dates for CODE. My dataframe contains prices from the start date to 08.06.2023.

For each code, I can calculate the returns for 30 - 60 - 90 days ahead of the code's own start date and find the average of those returns.

My start date is: 22.12.2022 (dd.mm.yyyy) for APPLE

My start date is: 16.11.2022 (dd.mm.yyyy) for HARLEY

My start date is: 21.10.2021 (dd.mm.yyyy) for META

DATE	CODE	CLOSE PRICE
08.06.2023	APPLE	8.5
07.06.2023	APPLE	7
06.06.2023	APPLE	9.3
05.06.2023	APPLE	12
04.06.2023	APPLE	3
......	APPLE	...
22.12.2022	APPLE	25
08.06.2023	HARLEY	4
07.06.2023	HARLEY	2
06.06.2023	HARLEY	1
05.06.2023	HARLEY	5
04.06.2023	HARLEY	9.5
...	HARLEY	....
16.11.2022	HARLEY	4
08.06.2023	META	1.1
07.06.2023	META	2
06.06.2023	META	1
05.06.2023	META	15
04.06.2023	META	65
...	META	...
21.10.2021	META	5

My expected output;

CODE	START DATE	AVERAGE RETURN FOR 30 DAY FROM START DATE
APPLE	22.12.2022	%2
HARLEY	16.11.2022	%4
META	21.10.2021	%-5

CODE	START DATE	AVERAGE RETURN FOR 60 DAY FROM START DATE
APPLE	22.12.2022	%6
HARLEY	16.11.2022	%3
META	21.10.2021	%-2

CODE	START DATE	AVERAGE RETURN FOR 90 DAY FROM START DATE
APPLE	22.12.2022	%-1
HARLEY	16.11.2022	%7
META	21.10.2021	%10

Solution

Even if your question is posed in the wrong manner, since you cannot expect that someone else do your homeworks from scratch in your place...
Here is my solution, from which you can take inspiration.

import pandas as pd
import numpy as np
from datetime import datetime, timedelta

data = {'DATE': ['08.06.2023', '07.06.2023', '06.06.2023', '05.06.2023', '04.06.2023', '22.12.2022', '08.06.2023', '07.06.2023', '06.06.2023', '05.06.2023', '04.06.2023', '16.11.2022', '08.06.2023', '07.06.2023','06.06.2023', '05.06.2023','04.06.2023', '21.10.2021'],
        'CODE': ['APPLE', 'APPLE', 'APPLE', 'APPLE', 'APPLE', 'APPLE', 'HARLEY', 'HARLEY', 'HARLEY', 'HARLEY', 'HARLEY', 'HARLEY', 'META', 'META', 'META', 'META', 'META', 'META'],
        'CLOSE PRICE': [8.5, 7, 9.3, 12, 3, 25, 4, 2, 1, 5, 9.5, 4, 1.1, 2, 1, 15, 65, 5]}

df = pd.DataFrame(data)

df['DATE'] = pd.to_datetime(df['DATE'], format='%d.%m.%Y')

start_dates = {'APPLE': datetime(2022, 12, 22),
               'HARLEY': datetime(2022, 11, 16),
               'META': datetime(2021, 10, 21)}
results = {}
for code in start_dates.keys():
    start_date = start_dates[code]
    portion_df = df.loc[df['CODE'] == code]
    if not portion_df.empty:
        prices = portion_df.loc[portion_df['DATE'] >= start_date, 'CLOSE PRICE'].values
        if len(prices) > 0:
            thirty_day_prices = prices[0:min(len(prices),30)]
            sixty_day_prices = prices[0:min(len(prices),60)]
            ninety_day_prices = prices[0:min(len(prices),90)]
            thirty_day_tot = (thirty_day_prices[-1] / thirty_day_prices[0]) - 1 if len(thirty_day_prices) > 1 else np.nan
            sixty_day_tot = (sixty_day_prices[-1] / sixty_day_prices[0]) - 1 if len(sixty_day_prices) > 1 else np.nan
            ninety_day_tot = (ninety_day_prices[-1] / ninety_day_prices[0]) - 1 if len(ninety_day_prices) > 1 else np.nan
            avg_thirty_day_return = np.nanmean(thirty_day_tot)
            avg_sixty_day_return = np.nanmean(sixty_day_tot)
            avg_ninety_day_return = np.nanmean(ninety_day_tot)
            results[code] = {'START DATE': start_date.strftime('%d.%m.%Y'),
                             'AVERAGE GAIN FOR 30 DAY FROM START DATE': f'{avg_thirty_day_return:.0%}' if not np.isnan(avg_thirty_day_return) else 'No data',
                             'AVERAGE GAIN FOR 60 DAY FROM START DATE': f'{avg_sixty_day_return:.0%}' if not np.isnan(avg_sixty_day_return) else 'No data',
                             'AVERAGE GAIN FOR 90 DAY FROM START DATE': f'{avg_ninety_day_return:.0%}' if not np.isnan(avg_ninety_day_return) else 'No data'}
        else:
            print(f"No data available for {code} with starting date {start_date.strftime('%d.%m.%Y')}")
    else:
        print(f"No data available for {code}")

wanted_df = pd.DataFrame.from_dict(results, orient='index')
wanted_df.index.name = 'CODE'

Hoping that it is what you wanted to achieve...Bye!