Bioregulators are substances applied to various types of crops to optimize their quantity and quality. Their external application could affect the plant metabolome by triggering metabolic changes and consequently modifying the quality of beverage. The present study investigated the metabolomic modifications observed in the spectral fingerprints of raw coffee beans dependent on bioregulator applications (BIO, Gibberellin, and Ethephon) over the stratified vertical canopy profile (STR). Principal components analysis (PCA) was used to identify the discriminating solvent, obtained from statistical mixture design, for each applied technique. When analyzing both the near-infrared (NIR) and UV–Visible (UV–Vis) spectra, E (ethanol) and ED (ethanol/dichloromethane) were the most discriminating solvents. The NIR and UV spectra obtained from the statistical mixture design were interpreted by using a modification of analysis of variance-common dimensions analysis (AComDim), where the usual PCA step within ComDim was replaced by an independent components analysis (ICA), i.e., AComDim-ICA. AComDim-ICA showed that ED extractor was the most efficient in bioregulator effect discrimination, while E was the most efficient for the stratification and the BIO x STR interactions. AComDim-ICA indicated that gibberellin gave the greatest differentiation compared to the other BIO treatments (Ethephon and Control), with a possible influence on phenolic compounds. AComDim-ICA indicated that berries harvested at strata up to 1 m (highly shaded) were different than berries harvested at higher strata (2 m and 3 m), that are related to metabolites such are caffeine, lipids, sugars, and diterpenes. Caffeine from the highest stratum (3 m) treated with Ethephon differed from that of the Control for the same stratum, in relation to caffeine. The use of these multivariate methods expands the metabolomic analysis of the plants, producing more information about the system.