Interpolation techniques are described for use with data that may not be uniform and may be characterized as scattered. Such data may be obtained in instances where data acquisition may not be easily controlled such as in obtaining experimental data for use with models. Data interpolation techniques may be used in connection with the experimental data to produce a more complete and accurate data set representative of a variety of conditions using as input the non-uniform or scattered data. Such data sets may be used in a variety of applications including providing a realistic and complete set of data for training and verifying neural networks.