Dataframe % column by grouping

Question

我正在制作一份预测准确性报告，该报告衡量实际预测与先前预测之间的偏差。测量结果为 = 1- ('Actual' - 'M-1') / 'Actual' 。需要根据不同的小费对措施进行分组，比如 'Product Category' / 'Line' / 'Product'。但是，df.groupby('Product Category').sum() 函数不支持百分比计算。有谁知道应该如何解决？谢谢！

data = {
    "Product Category": ['Drink', 'Drink','Drink','Food','Food','Food'],
    "Line": ['Water', 'Water','Wine','Fruit','Fruit','Fruit'],
    "Product": ['A', 'B', 'C','D','E','F'],
    "Actual": [100,50,40,20,70,50],
    "M-1": [120,40,10,20,80,50],
}
df = pd.DataFrame(data)

df['M1 Gap'] = df['Actual'] - df['M-1']
df['Error_Per'] =  1- df['M1 Gap'] / df['Actual']

预期输出为 enter image description here

Answer 1

您应该在计算百分比之前分组：

data = {
    "Product Category": ['Drink', 'Drink','Drink','Food','Food','Food'],
    "Line": ['Water', 'Water','Wine','Fruit','Fruit','Fruit'],
    "Product": ['A', 'B', 'C','D','E','F'],
    "Actual": [100,50,40,20,70,50],
    "M-1": [120,40,10,20,80,50],
}
df = pd.DataFrame(data)

df['M1 Gap'] = df['Actual'] - df['M-1']

df_line = df.groupby('Line').sum()
df_line['Error_Per'] =  df_line['M1 Gap'] / df_line['Actual']
print(df_line)

df_prod = df.groupby('Product Category').sum()
df_prod['Error_Per'] =  df_prod['M1 Gap'] / df_prod['Actual']
print(df_prod)

输出：

       Actual  M-1  M1 Gap  Error_Per
Line
Fruit     140  150     -10  -0.071429
Water     150  160     -10  -0.066667
Wine       40   10      30   0.750000

                  Actual  M-1  M1 Gap  Error_Per
Product Category
Drink                190  170      20   0.105263
Food                 140  150     -10  -0.071429

注意：您的屏幕截图预期结果与您的代码字典（我使用的）不匹配

Answer 2

您还可以创建一个自定义函数，并 apply 它在 pandas 数据框的每一行上，如下所示。请注意，我将 axis 参数设置为 1，以便自定义函数应用于每一行或跨列：

import pandas as pd

def func(row):
    row['M1 Gap'] = row['Actual'] - row['M-1']
    row['Error_Per'] = 1 - (row['M1 Gap'] / row['Actual'])
    return row


df.groupby('Product Category').sum().apply(func, axis = 1)


                  Actual    M-1  M1 Gap  Error_Per
Product Category                                  
Drink              190.0  170.0    20.0   0.894737
Food               140.0  150.0   -10.0   1.071429

Dataframe % column by grouping

Data Frame % column by groupping

dataframe

pandas

pandas-groupby