Creating image over non-rectangular grid

This is probably more of a help-wanted request than issue, but I wanted to visualize a large spectrogram. The spectrogram is plotted over a 2D non-rectangular grid (n_frequencies, n_timepoints) for a total of about 2 billion data points. For reference, here is my code (using matplotlib). I'm hoping to use datashader since my existing solution is very slow to render.

```
time, freq = np.meshgrid(timevec, freqvec)
# data is an array of shape (len(freqvec), len(timevec))
# whose values are the magnitude of the spectrogram
ax.contourf(time, freq, data)
```

I'm a newcomer and didn't find an example notebook that matched this application directly, so I'd definitely appreciate it if you could help or point me to the right resource. Thanks!

The issue provides a Python snippet using NumPy meshgrid, Matplotlib contourf, and spectrogram data, but names no repository file or test. Start by checking existing Datashader examples and APIs for non-rectangular grids; done would require a documented, reproducible path or an explicit decision about support.

Written by the indexing model from the issue text.