# Impute missing position data using nearby values?

Data looks like this:

time longitude latitude obs
1     xx          yy     zz
2     xx          yy     zz
3                        zz
4     xx          yy     zz


and so on. There are sometimes missing values for the longitude and latitude covariates. Since these are somewhat dependent on the longitudes and latitudes for "nearby" measurements ("nearby" with respect to "time"), I was thinking that I could just take the mean of such nearby measurements in order to impute these missing values.

Is it okay?