This invention includes models and software for real-time Raman spectral analysis of sample streams in traditional flow and static solution cells and microfluidic devices. It utilizes a database/training set and associated data treatment (developed and described under a separate disclosure) to identify and quantify species of interest, including UO22+, nitric acid, and total nitrate. This model and software is dependent on an associated database of spectra and its associated data treatment. The model produced uses COTS software in a Matlab environment (The MathWorks, Inc) using the PLS_Toolbox (Eigenvector Research Inc.). The associated database and training set is further treated for preparing the data for model construction. The unique property of the data treatment is choice of algorithm applied to the database/training set used for model construction. Examples of data treatment and algorithm choice for model development include 1) choice of non-linear vs linear modeling to deal with non-linear behavior; 2) use of archetypal binning of data to accommodate various concentration ranges.