基于通过散点的类别的颜色 3D 表面

问题描述

我有以下格式的数据:

X,Y,Z,类别

我使用 plotly 生成散点图，然后使用以下代码通过散点拟合曲线.

from scipy.interpolate import griddata将 numpy 导入为 np将 matplotlib.pyplot 导入为 plt从 matplotlib 导入厘米从 mpl_toolkits.mplot3d 导入 Axes3Dx=np.asarray([3,5,9,3,3,7,6,9,1,9]);y=np.asarray([4,3,3,10,8,2,4,10,9,3]);z=np.asarray([1,2,4,10,1,7,10,3,1,7]);# x = np.random.random(100)xi=np.linspace(min(x), max(x),50)#打印xiyi=np.linspace(min(y),max(y),50)X,Y= np.meshgrid(xi,yi)Z = np.nan_to_num(griddata((x,y), z, (X, Y), method='cubic'))无花果 = plt.figure()ax = fig.add_subplot(111, 投影='3d')ax.scatter(x, y, z)ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=cm.coolwarm,线宽=0，抗锯齿=假，alpha=0.4)plt.show()

我想做的是根据这样的类别为情节着色:其中红色代表类别 1，蓝色代表类别 2.所以为了得到这样的东西，我需要生成一个 2D 数组，然后使用颜色图/色标对类别进行相应的着色.

上面的输出是使用 XLSTAT 创建的，其中将类别作为第 4 列作为类别.

谁能解释我如何生成 Z 数据以帮助我对类别进行不同的着色?

我尝试过将二维矩阵分成两半 0 和半 1 并得到类似这样的输出.

考虑以下示例数据:

x y z 类别3 4 1 猫 15 3 2 类 29 3 4 类别 23 10 10 三类3 8 1 类 37 2 7 类别 26 4 10 类别 19 10 3 类别 41 9 1 猫 19 3 7 类别 2

我需要生成 2D 数据来表示表面颜色并使用自定义颜色为不同类别着色

解决方案

就像

I have the data in the following format:

X,Y,Z,Category

I used plotly to generate a scatter plot and then a fit a Curve through the scatter points using the following code.

from scipy.interpolate import griddata
import numpy as np
import matplotlib.pyplot as plt
from matplotlib import cm
from mpl_toolkits.mplot3d import Axes3D

x=np.asarray([3,5,9,3,3,7,6,9,1,9]);
y=np.asarray([4,3,3,10,8,2,4,10,9,3]);
z=np.asarray([1,2,4,10,1,7,10,3,1,7]);
# x = np.random.random(100)

xi=np.linspace(min(x), max(x),50)
#print xi
yi=np.linspace(min(y),max(y),50)


X,Y= np.meshgrid(xi,yi)
Z = np.nan_to_num(griddata((x,y), z, (X, Y), method='cubic'))

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(x, y, z)
ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=cm.coolwarm,
                       linewidth=0, antialiased=False,alpha=0.4)
plt.show()

What i am looking to do is to color the plot according to categories something like this : Where red represents the category 1 and Blue represents category 2. So inorder to get something like this I need to generate a 2D Array and then use a colormap/colorscale to color the categories accordingly.

The above output have been created using XLSTAT where it took category as the 4th col as the category.

Can someone explain me how do i generate the Z data which will help me color the categories differently?

I have tried to something like dividing the 2D matrix into halves 0's and half 1's and got output something like this.

Considering the following sample data :

x   y   z   Category
3   4   1   Cat 1
5   3   2   cat2
9   3   4   cat2
3   10  10  cat3
3   8   1   cat3
7   2   7   cat2
6   4   10  Cat 1
9   10  3   Cat 4
1   9   1   Cat 1
9   3   7   cat2

I need to generate 2D Data that will represent the surface color and color the different categories with custom color

解决方案

Just as griddata can be used to interpolate the 1D z array to a 2D grid, you can use griddata to interpolate a 1D color array to the same 2D grid:

color = [colormap[cat] for cat in category]
C = np.nan_to_num(griddata((x, y), color, (X, Y), method='cubic'))

Then you can use the colormap cm.coolwarm to map values in C to RGBA facecolors:

ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=cmap,
                linewidth=0, antialiased=False, alpha=0.4, facecolors=cmap(C))

---

import numpy as np
import matplotlib.pyplot as plt
from matplotlib import cm
from mpl_toolkits.mplot3d import Axes3D
import scipy.interpolate as interpolate

x = np.asarray([3, 5, 9, 3, 3, 7, 6, 9, 1, 9])
y = np.asarray([4, 3, 3, 10, 8, 2, 4, 10, 9, 3])
z = np.asarray([1, 2, 4, 10, 1, 7, 10, 3, 1, 7])
category = np.array(['Cat 1', 'cat2', 'cat2', 'cat3', 'cat3',
                     'cat2', 'Cat 1', 'Cat 4', 'Cat 1', 'cat2'])
# coolwarm: 0 --> blue, 1 --> red
# want: 'Cat 1' --> blue, 'cat2' --> red, 'cat3' --> ?, 'Cat 4' --> ?
colormap = {'Cat 1': 0, 'cat2': 1, 'cat3': 0.333, 'Cat 4': 0.666}
color = np.array([colormap[cat] for cat in category])

xi = np.linspace(min(x), max(x), 50)
yi = np.linspace(min(y), max(y), 50)
X, Y = np.meshgrid(xi, yi)
Z = np.nan_to_num(interpolate.griddata((x, y), z, (X, Y), method='cubic'))
C = interpolate.griddata((x, y), color, (X, Y), method='cubic')

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
cmap = cm.coolwarm
ax.scatter(x, y, z, c=color, cmap=cmap)
ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=cmap,
                linewidth=0, antialiased=False, alpha=0.4, facecolors=cmap(C))
plt.show()