Question

I am currently able to plot a pcolor style plot from some data. The data consists of three coordinates per data point (X,Y,Z).

My data is structured in the following way:

X: 1D array
   shape = [500,0]

Y: 1D array
   shape = [700,0]

Z: 2D array
   shape = [500, 700]

So the rows of my Z array correspond to the X coordinate, and my columns correspond to the Y coordinate.

Then to plot the data on a pcolor plot, I simply did:

plt.pcolormesh(X,Y,Z.T)

The transpose is needed because of how pcolormesh expects to receive the input data.

Now I want to plot the same data as a surface using plot_surface. I am having issues, however, in structuring my data the necessary way. Can anyone offer any guidance?

Answer 1

plot_surface expects all the inputs arrays to be 2D.

You can use numpy.meshgrid to generate the grid points

In [78]: X = np.arange(500)                 
In [79]: Y = np.arange(700)

In [83]: xx, yy = np.meshgrid(X, Y)

In [84]: xx.shape
Out[84]: (700, 500)

In [85]: yy.shape
Out[85]: (700, 500)

You can then call plot_surface(xx, yy, Z.T)