3D plot with matplotlib from imported data

I am looking to plot a 3D surface. The data is generated in mathematica and exported in a file with three columns separated by spaces (I can switch it to comma or other delimiter). The problem I have is that so far I have been using gnuplot for all my 2D plots, but for this particular plot gnuplot is not really that easy to use and the results are not that great.

I have read about matplotlib but I find the documentation really difficult to get through. When I started with gnuplot, I found a really good template here which I can use to generate 2D plots and multiplots any way I like.

Is there something similar for matplotlib that I can use? Almost all examples for matplotlib involve functions that are plotted in various ways but there aren't any examples where data is imported and the plotted.

1 Answer

I think you're asking several questions here, I will try and answer the most fundamental one: loading your CSV data into Python and then plotting a basic surface plot.

Note that in the general case you need an NxN matrix of Z values for a plot where X and Y are vectors of length N. This makes sense because every (x, y) point needs a Z value and there are N x N (x, y) points. I will show you two ways. Once you have those then at least you are in a good position to start customizing your plot using documentation and other examples found on the web.

This answer assumes you are using Python 3 and a recent version of Matplotlib and Numpy. (Numpy is a numerical Python package that is commonly included in Python distributions, e.g. Anaconda.) However, syntax should be not be very different in older versions.

First, make the necessary imports:

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

For the case when you have 3 columns of 1d data --- x, y and z:

DataAll1D = np.loadtxt("datacsv_1d.csv", delimiter=",")
# create 2d x,y grid (both X and Y will be 2d)
X, Y = np.meshgrid(DataAll1D[:,0], DataAll1D[:,1])
# repeat Z to make it a 2d grid
Z = np.tile(DataAll1D[:,2], (len(DataAll1D[:,2]), 1))
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.plot_surface(X, Y, Z, cmap='ocean')
plt.show()

Note than in np.loadtxt you can set any delimiter, e.g. ' ' for a space.

For the case when you have 2 columns of 1d data --- x and y --- and a 2d matrix of z data:

DataX_Y_1D = np.loadtxt("datacsv_1d_xy.csv", delimiter=",")
# create 2d x,y grid (both X and Y will be 2d)
X, Y = np.meshgrid(DataX_Y_1D[:,0], DataX_Y_1D[:,1])
# get 2D z data
Z = np.loadtxt("datacsv_2d_Z.csv", delimiter=",")
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.plot_surface(X, Y, Z, cmap='ocean')
plt.show()

For more advanced customization, here's a link to another good tutorial on the subject:

Edit: For the case where you have all the necessary data points covered in 3 columns (X, Y, Z) then plot_trisurf is what you need. See my answer on Stack Overflow for a simple example: