把我的评论作为答案：
您可以使用.unstack来实现这一点，顾名思义，它删除了MultiIndex的级别以创建多列索引：

df.unstack()

默认情况下，这将取消堆叠最后一个级别，但如果您希望使用level="AggregationMethod"、level=-1（默认值）或level=2来指定此选项，则可以指定此选项。

下面是一个如何使用Pandas pivot（如评论中所建议的）和reindex的示例：

import random

import pandas as pd

# Toy dataframe (same as yours after reseting the index)
df = pd.DataFrame(
    {
        "Entity Type": ["foo" for _ in range(20)],
        "Quantile Range": ["q99_q100", "q90_q99", "q70_q90", "q00_q70"] * 5,
        "AggregationMethod": ["max", "avg", "std", "total", "count"] * 4,
        "bar": [random.uniform(1, 999) for _ in range(20)],
        "baz": [random.uniform(1, 999) for _ in range(20)],
    }
)

df = df.pivot(
    index=["Entity Type", "Quantile Range"],
    columns="AggregationMethod",
    values=["bar", "baz"],
).reindex(["max", "avg", "std", "total"], level=1, axis=1)

print(df)
# Output

                                   bar
AggregationMethod                  max         avg         std       total   
Entity Type Quantile Range
foo         q00_q70         141.752307  822.270987  199.740853  595.444166  \
            q70_q90         383.574450  410.730888  838.562828  545.299705   
            q90_q99         339.588340  606.983173  935.142608  407.674059   
            q99_q100        161.833517  932.267262  157.149458  618.105967   

                                   baz
AggregationMethod                  max         avg         std       total  
Entity Type Quantile Range
foo         q00_q70          17.986766  298.760555  389.559554   49.925246  
            q70_q90         888.435092  695.713473  502.429534  209.356226  
            q90_q99         715.425998  209.749918  136.480141  525.729657  
            q99_q100        705.721265  956.273655  684.883477   39.114393