How to create gradient of randomly spaced data at each point of input array

I have a set of data, y, with non-uniformly spaced independent variable, x. I want to calculate the gradient of this data at each point in y so that dy/dx is the same length as y. If i use gradient, the data points in x must be equally spaced. If I use diff(y)./diff(x) the length of the resulting array is different than y. Suggestions?

Answer 1

[Star Strider]

For unequally-spaced points, I always use:

dydx = gradient(y)./gradient(x);

It depends on the result you want.

Answer 2

[Ameer Hamza]

If you have a model of the data (i.e. the relation between x and y is polynomial or exponential etc.), then you can use curve fitting to fit a curve to the data using fit() function. After a curve fit is obtained, you can find derivative at the specified points using differentiate().