<p>The present study reports a quantitative analysis of the gelation kinetics and mechanical behaviour of chitosan-based hydrogels chemically crosslinked with glyoxal. Time-resolved mechanical spectroscopy (TRMS) was coupled with non-parametric generalised additive models (GAM) to accurately model the evolution of viscoelastic moduli and determine statistically key parameters at the gel point, including gel time (t_g), gel stiffness (S), and relaxation exponent (Δ). The effect of polymer and crosslinker concentrations on these parameters was established with the use of GAM and validated by multivariate linear regression, yielding interpretable parametric models. In a similar way, a detailed statistical analysis of oscillatory and uniaxial compression measurements of fully gelled hydrogels in both swollen and dry states revealed distinct trends in mechanical response: crosslinker concentration enhanced stiffness in the swollen state but exhibited a saturation effect in the dry state. Convolutional neural network (U-Net) analysis of scanning electron microscopy (SEM) images was, finally, able to provide structural insight correlating pore morphology with mechanical properties. The findings demonstrate the utility of statistical learning tools in increasing knowledge of the mechanical and morphological properties of hydrogels.</p> Graphical Abstract <p></p>