python - 如何对 pandas DataFrame 进行分组和索引以获得所需的箱形图

Question

我的数据如下:

Region = [random.choice([1,2,3,4,5]) for x in range(100)]
Gender = [random.choice(['Male', 'Female']) for x in range(100)]
Balance = [random.random()*1000 for x in range(100)]
df = pd.DataFrame({'Region':Region, 'Gender':Gender, 'Balance':Balance})

我想获得一个带有索引(区域、性别)的多索引 DataFrame，以便我可以调用 df.plot.box(vert=False) 并获得类似的内容(在 R 中生成) )出来。

这看起来应该很简单，但我似乎无法获得正确的 reshape /索引。

Answer 1

np.random.seed(23)
Region = np.random.choice([1,2,3,4,5], size=100)
Gender = np.random.choice(['Male', 'Female'], size=100)
Balance = np.random.rand(100)*1000 
df = pd.DataFrame({'Region':Region, 'Gender':Gender, 'Balance':Balance})
print (df.head())
      Balance  Gender  Region
0  384.491355  Female       4
1  328.787350  Female       1
2  529.003182    Male       2
3   96.884964  Female       1
4   23.379931    Male       5

我认为需要首先连接Region和Gender，使用cumcount用于对pivot的每个组进行计数:

idx = df['Region'].astype(str) + '.' + df['Gender']
cols = idx.groupby(idx).cumcount()
df1 = pd.pivot(index=cols, columns=idx, values=df['Balance'])
print (df1)

print (df1.head())

     1.Female      1.Male    2.Female      2.Male    3.Female      3.Male  \
0  328.787350  298.232904  888.262152  529.003182  959.644810  962.342645   
1   96.884964  780.852785  738.040024  760.956146  119.652522  601.118950   
2  910.707827  611.333680  116.517822  155.214746  140.653479  688.654958   
3   50.119030  205.932674  148.848025  794.379306  380.307363  194.257663   
4  263.554386  605.087006  953.241083  113.801236  778.912082  170.791317   

     4.Female      4.Male    5.Female      5.Male  
0  384.491355  122.347230  400.107360   23.379931  
1  190.038651  564.785449  330.269653  998.586681  
2  521.390446  757.714947  512.813561  185.192917  
3  566.314099  939.538858  480.686727   80.862220  
4  927.260017  175.496721  342.465179  287.932951  

df1.plot.box(vert=False)

<小时/>

旧的解决方案:

看来你需要通过groupby reshape 并聚合平均值或pivot_table :

a = df.groupby(['Gender','Region'])['Balance'].mean().unstack()
#alternatively
#a = df.pivot_table(index='Gender', columns='Region', values='Balance', aggfunc='mean')
print (a)
Region           1           2           3           4           5
Gender                                                            
Female  357.970914  679.143664  442.473514  498.600391  618.475656
Male    531.211030  462.071729  470.280364  623.540595  362.917609

a.plot.box(vert=False)


b = df.groupby(['Region','Gender'])['Balance'].mean().unstack()
print (b)
Gender      Female        Male
Region                        
1       357.970914  531.211030
2       679.143664  462.071729
3       442.473514  470.280364
4       498.600391  623.540595
5       618.475656  362.917609

b.plot.box(vert=False)