带有极限点和异常点的极坐标图

Question

考虑以下数据框，

d = {'Score': [0.25, 0.52, 0.26, 0.22, 0.31, 2.45, 3.68, 41.3, 87, 91], 
     'Thr1': 16.5,
     'Thr2': 45.5,  
     'Anomaly':[0, 0, 0, 0, 0, 0, 0, 1, 1, 1]}

df = pd.DataFrame(data = d)

我要做的是绘制一个极坐标图，用虚线表示阈值，或用多条虚线表示多个阈值，并用不同的颜色表示异常。到目前为止，我得到的是，

r = df['Score']
theta = df.index.values
fig = plt.figure()
ax = fig.add_subplot(111, projection = 'polar')
c = ax.scatter(theta, r)

我无法获得阈值并更改异常点的颜色。有什么想法吗？

Answer 1

您需要在阈值级别绘制一条虚线，以指示阈值在哪里。 （一条线将在极坐标图上显示为一个圆圈）。

然后您需要根据它们是否低于、介于或高于阈值来分离要绘制在散点图上的值，并相应地为点着色。

import matplotlib
matplotlib.use('TkAgg')

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


dataset = {'Score': [0.25, 0.52, 0.26, 0.22, 0.31, 2.45, 3.68, 41.3, 87, 91], 
           'Thr1': 16.5,
           'Thr2': 45.5,  
           'Anomaly':[0, 0, 0, 0, 0, 0, 0, 1, 1, 1]}

df = pd.DataFrame(data=dataset)

scores = df['Score']
theta, thr_1, thr_2 = df.index.values, dataset['Thr1'], dataset['Thr2']

fig = plt.figure()
ax = fig.add_subplot(111, projection='polar')

# assigns a color to each point based on their relative value to the thresholds
colors = ['b' if val < thr_1 else 'y' if val < thr_2 else 'r' for val in scores]
point_cloud = ax.scatter(theta, scores, color=colors, marker='o')

# Drawing the threshold dash lines (with alpha value 1/2)
theta_xs, thr_y1, thr_y2 = np.linspace(0, 2*np.pi, 20), [thr_1] * 20, [thr_2] * 20
thr_line_1 = ax.plot(theta_xs, thr_y1, color='blue', linestyle='--', alpha=0.5)
thr_line_2 = ax.plot(theta_xs, thr_y2, color='green', linestyle='--', alpha=0.5)

plt.show()

Answer 2

好吧，我不确定这是否是您想要的，因为我从未使用过您的数据集的 Anomaly 部分，而只是从 Score 数组中获取颜色信息

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.colors as c
d = {'Score': [0.25, 0.52, 0.26, 0.22, 0.31, 2.45, 3.68, 41.3, 87, 91],
     'Thr1': 16.5,
     'Thr2': 45.5,
     'Anomaly': [0, 0, 0, 0, 0, 0, 0, 1, 1, 1]}

df = pd.DataFrame(data = d)

r = df['Score']
theta = df.index.values
fig = plt.figure()
ax = fig.add_subplot(111, projection = 'polar')

#Add thresholds
ax.plot(np.linspace(0, 2*np.pi, 100), np.ones(100)*d['Thr1'], c='g', ls='--')
ax.plot(np.linspace(0, 2*np.pi, 100), np.ones(100)*d['Thr2'], c='r', ls='--')

#Add colors
colors = ['g' if v < d['Thr1'] else 'y' if v < d['Thr2'] else "r" for v in r]
sc = ax.scatter(theta, r, c=colors)

plt.show()