How to produce a conditional probability distribution of P( X>x | Y>y )

Hello all,

I have my data pair as a 20x2 matrix. Lets call the first column to be my X vector and second as my Y vector, which are my actual data set. X and Y are partially correlated on a log-log space (corr. coeff=0.7). I want to derive a lognormal cumulative conditional distribution of the following: P( X>x | Y>y ) where lower case letters denote limit values. For a single plot I want the X-axis of my plot to cover the possible ranges that 'y' can take. I cant figure out how to produce this conditional distribution I just mentioned in Matlab. Any sort of help is greatly appreciated. Thanks!