We presented and tested a new method of interpolation of tracked datasets with non-Gaussian errors onto a regular grid. We showed that with the Galilee dataset, our method was able to remove artifacts caused by irregular spikes in the data as well as the acquisition footprint caused by systematic errors in the data. We think that this method is directly applicable to other problems where measurements are taken along the tracks and data have errors due to equipment malfunctioning and systematic errors due to slowly changing acquisition conditions.