How to plot a smooth 2D color plot for z = f(x, y)

Question

I am trying to plot 2D field data using matplotlib. So basically I want something similar to this:

enter image description here

In my actual case I have data stored in a file on my harddrive. However for simplicity consider the function z = f(x, y). I want a smooth 2D plot where z is visualised using color. I managed the plotting with the following lines of code:

import numpy as np
import matplotlib.pyplot as plt

x = np.linspace(-1, 1, 21)
y = np.linspace(-1, 1, 21)
z = np.array([i*i+j*j for j in y for i in x])

X, Y = np.meshgrid(x, y)
Z = z.reshape(21, 21)

plt.pcolor(X, Y, Z)
plt.show()

However, the plot I obtain is very coarse. Is there a very simple way to smooth the plot? I know something similar is possible with surface plots, however, those are 3D. I could change the camera angle to obtain a 2D representation, but I am convinced there is an easier way. I also tried imshow but then I have to think in graphic coordinates where the origin is in the upper left corner.

Problem solved

I managed to solve my problem using:

plt.imshow(Z,origin='lower',interpolation='bilinear')

The only thing that separates the linked plot with yours is that your grid is much coarser. Change 21 to something larger, say 101, or 1001. — wflynny, May 28 '15 at 14:56
Yes I know, but even with a low number of grid points a smooth plot can be obtained using interpolation. `imshow` has that ability. Can I do something similar with `pcolor`? — Aeronaelius, May 28 '15 at 14:58
Please edit your question to describe _exactly_ what you want, with a representative example. — wflynny, May 28 '15 at 15:02

wflynny · Accepted Answer · 2015-05-28T15:07:07.120

12

If you can't change your mesh granularity, then try to go with imshow, which will essentially plot any 2D matrix as an image, where the values of each matrix cell represent the color to make that pixel. Using your example values:

In [3]: x = y = np.linspace(-1, 1, 21)
In [4]: z = np.array([i*i+j*j for j in y for i in x])
In [5]: Z = z.reshape(21, 21)
In [7]: plt.imshow(Z, interpolation='bilinear')
Out[7]: <matplotlib.image.AxesImage at 0x7f4864277650>
In [8]: plt.show()

enter image description here

edited May 28 '15 at 15:07

answered May 28 '15 at 14:59

wflynny

18,065
5
46
67

I am aware of that. But the case I presented was a simple one. I actually have data stored in a `txt` file so I cannot adjust the data. So yes, in a simple function case I could do that. But what do you do in case of data obtained from a 3rd party? – Aeronaelius May 28 '15 at 15:00
1

Thanks I came to this too just now. I also added `origin='lower'` as a parameter to have the origin at the lower left corner. – Aeronaelius May 28 '15 at 15:17

tmdavison · Answer 2 · 2015-05-28T15:14:32.507

12

you can use contourf

plt.contourf(X, Y, Z)

enter image description here

EDIT:

For more levels (smoother colour transitions), you can use more levels (contours)

For example:

plt.contourf(X, Y, Z, 100)

enter image description here

edited May 28 '15 at 15:14

answered May 28 '15 at 15:06

tmdavison

64,360
12
187
165

1

This has clear separation between regions. I want a smooth gradient – Aeronaelius May 28 '15 at 15:08
2

you can increase the number of levels. I'll update with instructions – tmdavison May 28 '15 at 15:13

score 0 · Answer 3 · answered Aug 10 '23 at 09:10

You can do this using pcolor or pcolormesh if you provide an option to use Gouraud shading:

import numpy as np
from matplotlib import pyplot as plt

y, x = np.meshgrid(np.linspace(-1, 1, 5), np.linspace(-1, 1, 5))
z = x**2+y**2
_, [ax1, ax2] = plt.subplots(1,2)
ax1.pcolormesh(x, y, z)
ax1.set_title("Default shading")
ax2.pcolormesh(x, y, z, shading='gouraud')
ax2.set_title("Gouraud shaing")

How to plot a smooth 2D color plot for z = f(x, y)

3 Answers3

Linked