Removing NaN's and interpolating missing data

Hi all

I have a small problem with some data I'm trying to analyse.

I have some spreadsheet data which, in excel looks like this: (Google docs spreadsheet).

https://docs.google.com/spreadsheet/ccc?key=0AliwBByzNrWtdGUxVDQtZUlWOUpKV2s0MGRadjd0c0E

When I load this into MATLAB, it replaces the blanks with NaNs

What I need to do is remove the NaNs and replace them with values interpolated or extrapolated from the values around them, so that I can produce a surface plot using countourf. The first row and first column are the axes - they are horizontal and vertical positions respectively.

I'm a relative beginner to MATLAB, so although I've found some previous posts about how to replace NaNs and interpolate, I'm having some difficulty adapting those solutions to my specific problem.

Any help greatly appreciated.

Thanks Steve

Answer 1

[Titus Edelhofer]

Hi,

a good starting point could be the File Exchange, e.g. this submission.

Titus

Answer 2

[Grzegorz Knor]

Take a look at this file:

http://www.mathworks.com/matlabcentral/fileexchange/4551

[Steve]

Hi thanks, that seems to be what I need, but I'm a little unsure about how to use it - excuse my ignorance, I'm just unfamiliar with the intricacies of how MATLAB works.

Do I need to modify the inpaint_nans.m code with the names of my own workspace variables or is there a more general way to use this? Having to edit the code each time would be painstaking (I have lots of these).

Thanks for any help

Steve