Data modeling, sometimes referred to as curve fitting, is of interest for one or both of two general reasons: (1) to allow an array of data (i.e., a table of data) to be represented by an analytical function, and (2) to allow the interpolation and/or extrapolation of those data. In either case, it must be decided whether the data need to be represented with a function that must represent or reproduce each of the original data points exactly, or if the data are to be represented by a smoothed curve where each data point is not necessarily represented exactly, but where the general trend of the data is of primary significance. We can model multi-dimensional data with a mathematical function that can be used to interpolate or extrapolate the given data. Extrapolation should always be done with caution, however, as obtaining data beyond the range of the data used for the model can yield unreasonable or unrealistic results.