Advanced preprocessing and analysis techniques for enhanced Raman spectroscopy data interpretation

Raman spectroscopy is a powerful, non-destructive technique for probing molecular composition. However, raw Raman spectra often include cosmic ray spikes, baseline drift from fluorescence, and other noise sources that hinder accurate interpretation. We present a comprehensive workflow combining cosmic ray removal (modified Zscore), data averaging, Savitzky-Golay smoothing, baseline correction (ALS/IARPLS), and iterative Voigt peak fitting. Using both synthetic and real-world data (Teflon, orthodontics aligners, and medical ointments), we demonstrate the robustness of our automated approach in enhancing signal-to-noise ratio (SNR) and improving the accuracy of peak identification. We discuss parameter selection, highlight the challenges posed by highfluorescence samples, and provide open-source code for reproducibility.