Cálculo do ponto de controle no Pandas

Para levar em conta os volumes, você precisa agrupar closee calcular a soma:

import pandas as pd

data = {
    'timestamp': ["2024-02-05 01:00:01.383985+00:00"]*18 + ["2024-02-05 01:00:01.385840+00:00"]*4,
    'close': [4968.5]*22,
    'security_code': ['ES']*22,
    'volume': [1]*18 + [2, 1, 1, 1],
    'bid_volume': [1]*18 + [0, 0, 0, 0],
    'ask_volume': [0]*18 + [2, 1, 1, 1]
}

df = pd.DataFrame(data)

def calculate_poc(df):
    volume_by_price = df.groupby('close')['volume'].sum()
    poc = volume_by_price.idxmax()
    return poc


poc_value = calculate_poc(df)
print(f"The Point of Control (POC) is at price: {poc_value}")

que retorna

The Point of Control (POC) is at price: 4968.5

Atualização: Adicionando uma janela contínua

Para fazer isso, você precisará determinar o menor horário de início e o maior horário de término e se preparar para percorrer esse intervalo em blocos (janelas) de x minutos cada (aqui escolhi 5). Em cada loop, calcule o preço de fechamento e some todos os volumes para cada preço.

import pandas as pd

# Data setup
data = {
    'timestamp': ["2024-02-05 01:00:01.383985+00:00"]*18 + ["2024-02-05 01:00:01.385840+00:00"]*4,
    'close': [4968.5]*22,
    'security_code': ['ES']*22,
    'volume': [1]*18 + [2, 1, 1, 1],
    'bid_volume': [1]*18 + [0]*4,
    'ask_volume': [0]*18 + [2, 1, 1, 1]
}
df = pd.DataFrame(data)
df['timestamp'] = pd.to_datetime(df['timestamp'])
df.set_index('timestamp', inplace=True)

def manual_rolling_poc(df, window='5T'):

    window_size = pd.to_timedelta(window)
    start_time = df.index.min()
    end_time = df.index.max()
    
    results = pd.Series(index=df.index, dtype='float64')  
    current_end_time = start_time + window_size

    while current_end_time <= end_time + window_size:
        window_df = df[(df.index >= start_time) & (df.index < current_end_time)]
        
        if not window_df.empty:
            volume_by_price = window_df.groupby('close')['volume'].sum()
            poc = volume_by_price.idxmax()
            results[(df.index >= start_time) & (df.index < current_end_time)] = poc
        else:
            results[(df.index >= start_time) & (df.index < current_end_time)] = None
        
        start_time += window_size
        current_end_time += window_size

    return results

df['rolling_poc'] = manual_rolling_poc(df)
print(df[['close', 'volume', 'rolling_poc']])

que vai dar

                                   close  volume  rolling_poc
timestamp                                                    
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.383985+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.385840+00:00  4968.5       2       4968.5
2024-02-05 01:00:01.385840+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.385840+00:00  4968.5       1       4968.5
2024-02-05 01:00:01.385840+00:00  4968.5       1       4968.5