The Sun exhibits variability over a wide range of timescales, from days to centuries, with long-term variations driven by the solar dynamo through the interplay of differential rotation, convection, and magnetic fields. While long-term photospheric variability has been extensively studied, the chromosphere, an essential interface linking the solar interior to the outer atmosphere, has remained comparatively underexplored over multidecadal timescales. The primary goal of this thesis is to investigate the long-term behaviour of the solar chromosphere and its coupling with the photosphere using century-long Ca II K observations. Using newly calibrated Ca ii K spectroheliograms from Kodaikanal Solar Observatory (1907–2007), along with complementary datasets from Meudon, Mt. Wilson, Rome/PSPT, and photospheric continuum data from MDI, we apply automated techniques to study chromospheric differential rotation. We find that chromospheric plages rotate faster than photospheric sunspots, corresponding to a 1.59% faster equatorial rotation and a weaker latitudinal shear, with no significant secular trend or hemispheric asymmetry over the past century. Using automated detection of Ca II K polar network features, we reconstruct the solar polar magnetic field over a 119-year period and demonstrate strong agreement with Wilcox Solar Observatory measurements and facular proxies, confirming the polar field’s predictive relationship with subsequent solar cycle amplitudes. Finally, we develop a convolutional neural network–based framework to extract solar disk and plage information from hand-drawn KoSO suncharts, successfully recovering plage areas and filling post-1980 data gaps in the Ca II K archive. Together, these results highlight the unique value of chromospheric archives for understanding long-term solar variability and provide new observational constraints for solar dynamo models, while demonstrating how modern automated and ML techniques can revitalise historical solar datasets.