Python 曲面绘图

Question

我有以下table数据（请看图片）为此，我想在 python 中绘制曲面图。使用来自 matplotlib、

的表面绘图

import matplotlib.pyplot as plt
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
import numpy as np

fig = plt.figure()
ax = fig.gca(projection='3d')
X=[2,3,5,8,20,30,50,80,100,150,175,200,250,300]
Y=[2,3,4,5,10,15,20,30,40,50,80,100,125,150,175,200]
Y,X=np.meshgrid(Y,X)
Z=np.array([
[0.2885307,0.269452,0.259193,0.2548041,0.2731868,0.4801551,0.7992361,1.7577641,3.2611327,5.428839,19.647976,37.59729,78.0871,152.21466,268.14572,0],
[0.2677955,0.2538363,0.2380033,0.2306999,0.4779794,0.9251045,1.5448972,3.508644,6.4968576,11.252151,0,0,0,0,0,0],
[0.2432982,0.2283371,0.2514196,0.3392502,0,0,0,0,0,0,0,0,0,0,0,0],
[0.2342575,0.3158406,0.4770729, 0.6795485,2.353042, 5.260077,9.78172,25.87004,59.52568, 0,0,0,0,0,0,0],
[0.6735384, 1.3873291,2.346506, 3.5654,0,0,0,0,0,0,0,0,0,0,0,0],
[1.3584715, 2.9405127,5.096819,8.155857,0,0,0,0,0,0,0,0,0,0,0,0],
[3.558062,8.216592,15.768077,27.386694,0,0,0,0,0,0,0,0,0,0,0,0],
[9.537899,25.202589,58.20041,0,0,0,0,0,0,0,0,0,0,0,0,0],
[16.083374,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[54.936775,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[89.185974,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]])

my_col = cm.jet(Z/np.amax(Z))
surf = ax.plot_surface(X, Y, Z,cmap=cm.coolwarm,linewidth=0, antialiased=False)
ax.set_zlim(0, 300)
ax.zaxis.set_major_locator(LinearLocator(10))
ax.zaxis.set_major_formatter(FormatStrFormatter('%.02f'))
fig.colorbar(surf, shrink=0.5, aspect=5)
plt.show()

我得到这样的情节

这是正确的，但不是很有吸引力或直观。如何使可视化更加流畅和清晰？请注意，我的数据中有很多空白。我应该使用“zero”作为空格还是“nan”（不是数字）？对于相同的数据，excel 显示的图表要好得多。我感谢您的意见，以使 python 情节更具视觉吸引力。

Answer 1

matplotlib 和 excel 绘图之间的区别在于 matplotlib 是在线性尺度上绘图而 excel 是对数的（或者看起来像对数轴但实际上不是的东西 - - 见下文）。因此，在 matplotlib 中，斜坡看起来非常陡峭，但在 excel 中，斜坡被日志显着拉长了。

不幸的是，matplotlib 还没有在 3D 中运行良好的对数轴。我不确定这是为什么，但这是一个严重的缺点。如果在绘制之前获取 X 和 Y 数据的 log10，则可以看到类似于 Excel 的图。您还可以进一步 DIY 自己的日志轴，但我刚刚使用刻度格式化程序为此完成了 shorthand。

import matplotlib.pyplot as plt
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter, FuncFormatter
from mpl_toolkits.mplot3d import axes3d
import numpy as np

def format_log(x, pos=None):
    x1 = 10**x
    s = "%.3f" % x1
    return s[:-4] if s[-3:]=="000" else " "

fig = plt.figure()
ax = fig.gca(projection='3d')
X=[2,3,5,8,20,30,50,80,100,150,175,200,250,300]
Y=[2,3,4,5,10,15,20,30,40,50,80,100,125,150,175,200]
X = np.log10(np.array(X))
Y = np.log10(np.array(Y))
Y,X=np.meshgrid(Y,X)
Z=np.array([
[0.2885307,0.269452,0.259193,0.2548041,0.2731868,0.4801551,0.7992361,1.7577641,3.2611327,5.428839,19.647976,37.59729,78.0871,152.21466,268.14572,0],
[0.2677955,0.2538363,0.2380033,0.2306999,0.4779794,0.9251045,1.5448972,3.508644,6.4968576,11.252151,0,0,0,0,0,0],
[0.2432982,0.2283371,0.2514196,0.3392502,0,0,0,0,0,0,0,0,0,0,0,0],
[0.2342575,0.3158406,0.4770729, 0.6795485,2.353042, 5.260077,9.78172,25.87004,59.52568, 0,0,0,0,0,0,0],
[0.6735384, 1.3873291,2.346506, 3.5654,0,0,0,0,0,0,0,0,0,0,0,0],
[1.3584715, 2.9405127,5.096819,8.155857,0,0,0,0,0,0,0,0,0,0,0,0],
[3.558062,8.216592,15.768077,27.386694,0,0,0,0,0,0,0,0,0,0,0,0],
[9.537899,25.202589,58.20041,0,0,0,0,0,0,0,0,0,0,0,0,0],
[16.083374,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[54.936775,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[89.185974,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]])
my_col = cm.jet(Z/np.amax(Z))
surf = ax.plot_surface(X, Y, Z,cmap=cm.coolwarm)
ax.set_zlim(0, 300)
ax.zaxis.set_major_locator(LinearLocator(10))
ax.zaxis.set_major_formatter(FormatStrFormatter('%.02f'))
ax.xaxis.set_major_formatter(FuncFormatter(format_log))
ax.yaxis.set_major_formatter(FuncFormatter(format_log))
fig.colorbar(surf, shrink=0.5, aspect=5)
plt.show()

编辑：
回到这个问题后，我意识到 Excel 图实际上并没有显示对数轴，而只是绘制给定 的 X 和 Y 值沿轴间距相等 即使这些值也没有明确的数学级数。

请务必注意，这不是数据的良好表示，因为它给人的明显印象是它是对数的（对于所提供的特定数据），但实际上并非如此，尽管它需要异常仔细地检查看到那个。这里相邻数字之间的差距甚至不是单调的。

所以我不鼓励这种表示，但为了重现 Excel 情节，我建议同样制作 spaced 数据，但用不同的数字标记它（仅这句话应该是足以阻止这种方法）。但这是代码和方法：

fig = plt.figure()
ax = fig.gca(projection='3d')
x=[2,3,5,8,20,30,50,80,100,150,175,200,250,300]
y=[2,3,4,5,10,15,20,30,40,50,80,100,125,150,175,200]

Y,X=np.meshgrid(range(len(y)),range(len(x)))
Z=np.array([
[0.2885307,0.269452,0.259193,0.2548041,0.2731868,0.4801551,0.7992361,1.7577641,3.2611327,5.428839,19.647976,37.59729,78.0871,152.21466,268.14572,0],
[0.2677955,0.2538363,0.2380033,0.2306999,0.4779794,0.9251045,1.5448972,3.508644,6.4968576,11.252151,0,0,0,0,0,0],
[0.2432982,0.2283371,0.2514196,0.3392502,0,0,0,0,0,0,0,0,0,0,0,0],
[0.2342575,0.3158406,0.4770729, 0.6795485,2.353042, 5.260077,9.78172,25.87004,59.52568, 0,0,0,0,0,0,0],
[0.6735384, 1.3873291,2.346506, 3.5654,0,0,0,0,0,0,0,0,0,0,0,0],
[1.3584715, 2.9405127,5.096819,8.155857,0,0,0,0,0,0,0,0,0,0,0,0],
[3.558062,8.216592,15.768077,27.386694,0,0,0,0,0,0,0,0,0,0,0,0],
[9.537899,25.202589,58.20041,0,0,0,0,0,0,0,0,0,0,0,0,0],
[16.083374,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[54.936775,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[89.185974,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]])
my_col = cm.jet(Z/np.amax(Z))
surf = ax.plot_surface(X, Y, Z,cmap=cm.coolwarm)
ax.tick_params(axis='both', which='major', labelsize=6)
ax.set_zlim(0, 300)
ax.xaxis.set_major_locator(IndexLocator(1, 0))
ax.xaxis.set_major_formatter(FixedFormatter([repr(v) for v in x]))
ax.yaxis.set_major_locator(IndexLocator(1, 0))
ax.yaxis.set_major_formatter(FixedFormatter([repr(v) for v in y]))
fig.colorbar(surf, shrink=0.5, aspect=5)

如果想显示 X 和 Y 的具体数字，一种解决方案是用对数轴绘制（因为数字非常接近 spaced以对数方式），然后在轴上专门用线绘制数字，或者，不要使用这些数字而不是通常的 spaced 数字。（但是要将这些绘制为轴值，和 space 以固定的间隔在视觉上显示它们，这是一个问题。）

Python 曲面绘图

Python surface plotting

python

excel

matplotlib

mplot3d