如何用andrew_curves制作熊猫资料？

Question

我有以下熊猫资料：

df = pd.read_csv('path/file/file.csv',
                 header=0, sep=',', names=['PhraseId', 'SentenceId', 'Phrase', 'Sentiment'])

我想用曲线打印它，我尝试了以下几种方法：

andrews_curves(df, 'Name')

知道怎么画这个吗？这是csv的内容：

PhraseId, SentenceId, Phrase, Sentiment
1, 1, A series of escapades demonstrating the adage that what is good for the goose is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story ., 1
2, 1, A series of escapades demonstrating the adage that what is good for the goose, 2
3, 1, A series, 2
4, 1, A, 2
5, 1, series, 2
6, 1, of escapades demonstrating the adage that what is good for the goose, 2
7, 1, of, 2
8, 1, escapades demonstrating the adage that what is good for the goose, 2
9, 1, escapades, 2
10, 1, demonstrating the adage that what is good for the goose, 2
11, 1, demonstrating the adage, 2
12, 1, demonstrating, 2
13, 1, the adage, 2
14, 1, the, 2
15, 1, adage, 2
16, 1, that what is good for the goose, 2
17, 1, that, 2
18, 1, what is good for the goose, 2
19, 1, what, 2
20, 1, is good for the goose, 2
21, 1, is, 2
22, 1, good for the goose, 3
23, 1, good, 3
24, 1, for the goose, 2
25, 1, for, 2
26, 1, the goose, 2
27, 1, goose, 2
28, 1, is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story ., 2
29, 1, is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story, 2

Answer 1

在链接到的文档页中，Iris数据集有一个名为'Name'的列。当你打电话的时候

andrews_curves(data, 'Name')

data的行按Name的值进行分组。这就是为什么对于Iris数据集来说，线条有三种不同的颜色。

在您的数据集中有三列：A、B、C。要在andrews_curves上调用df，首先需要标识要按其分组的值。例如，如果它是C列的值，则调用

andrews_curves(data, 'C')

另一方面，如果您想按列名、A、B、C进行分组，那么首先熔化您的DataFrame，将其从宽格式转换为长格式，然后在variable列上调用andrews_curves (该列保存每一行的值A、B或C )：

import numpy as np
import pandas as pd
import pandas.plotting as pdplt
import matplotlib.pyplot as plt

x = np.linspace(-1, 1, 1000)
df = pd.DataFrame({'A': np.sin(x**2)/x,
                   'B': np.sin(x)*np.exp(-x),
                   'C': np.cos(x)*x})
pdplt.andrews_curves(pd.melt(df), 'variable')
plt.show()

收益率

​

​