Scalings are the basis for relating and comparing quantities from scientific data. The current approach to discovering scalings from data is overly reliant on human researchers, which means that the pace of discovery in engineering and scientific fields is dependent on human speed. We propose a combined objective for quantitatively comparing scalings which will quicken the pace of discovery by harnessing computational speed. The combined objective allows for automated methods to be used to generate a large number (thousands or more) of candidates which can be computationally scored based on metrics of correlation, simplicity, sparsity, and having simple rational numbers. This method allows a researcher to explore pools of scaling candidates more efficiently to find scalings of interest.