Interpolating point data using scipy/python is not covering rectangular extent

Question

I am trying to write a python script with help of GDAL to a red point shapefile and considering a field value as z vales where I need interpolation to get raster with interpolated values. I am using scipy grid data for interpolation.

My interpolation code is:

zi = scipy.griddata((x, y), z, (xi, yi), method='cubic')

then I tried to write the interpolated data to raster and I got the following result as screen shot: enter image description here

It shows interpolation inside points. But, I want to interpolate my points to cover the whole rectangular space.

Does anyone have idea how it works?

Answer 1

The function is interpolating only inside the convex hull of data points https://docs.scipy.org/doc/scipy/reference/generated/scipy.interpolate.griddata.html.

Here smoothing results from scipy griddata interpolatioin is an example about how to add extrapolation but I do not understand why to do so because extrapolated values are meaningless. But if you want, I suppose you can also fool the algorithm easily by adding fake points to the corners of your area of interest.