带有熊猫和Jupyter笔记本的交互式箱线图

我在pandas DataFrame中有数据，我想创建一个交互式箱线图，让我可以选择天数，同时为“类别”列中每个类别的值绘制箱线图。

到目前为止，这是我的代码/数据的样子：

import numpy as np
import pandas as pd

categories=('A','B','C')

data = {
            'days':      np.random.randint(120, size=100), 
            'category':  np.random.choice(categories, 100),
            'value':     100.0 * np.random.random_sample(100)
       }

df = pd.DataFrame(data)

print(df)

   category  days      value
0         A     4  77.383981
1         A    31  63.011934
2         A     5   1.165061
3         C    59  23.588979
4         A    57  14.906734
5         C   106  33.366634
6         A    29  90.658570
7         B    25  16.137490
8         A   118  34.526302
9         C    76   4.111797
10        A    11  30.195917
..      ...   ...        ...
90        A    64  37.529774
91        A    76   3.771360
92        C   112  93.948775
93        C    14  34.855189
94        B    64  83.106007
95        A    10  78.346319
96        B    86  66.645889
97        A    46  12.969012
98        C    29  57.925427
99        A    59  34.526146

[100 rows x 3 columns]

我想为每个类别（针对选定的/指定的天数）创建值的箱形图，并沿X轴绘制不同的类别。

如何使用熊猫（或matplotlib）做到这一点？

您只需按天数过滤数据框，然后绘制相应的箱线图。

numer_of_days = 42
df_filtered= df.loc[df['days'] < numer_of_days]  # use operators like ==, >=, <, etc.
df_filtered[["category", "value"]].boxplot( by="category", return_type='axes')

import numpy as np
import pandas as pd
from ipywidgets import interact
%matplotlib notebook

categories=('A','B','C')

data = {
            'days':      np.random.randint(12, size=100), 
            'category':  np.random.choice(categories, 100),
            'value':     100.0 * np.random.random_sample(100)
       }

df = pd.DataFrame(data)

def select_days(number_of_days):
    df_filtered= df.loc[df['days'] == int(number_of_days)] 
    ax = df_filtered[["category", "value"]].boxplot( by="category", return_type='axes')
    ax["value"].set_title("Day " + number_of_days)
    print df_filtered

days = [str(day) for day in np.arange(12)]

interact(select_days, number_of_days=days)