||Spicules are ubiquitous, short-lived and highly dynamic features that permeate the structure of the magnetic solar chromosphere. They are broadly divided into two categories: type-I and type-II. Type-II spicules are the more energetic of the two types and are believed to impact the mass-loading of the solar corona. In this work we characterize, for the first time, type-II spicules in Ca II K 3934 A using an unsupervised machine learning technique, from the observations of the newly installed CHROMospheric Imaging Spectrometer (CHROMIS) at the Swedish 1-m Solar Telescope (SST). We find that their line formation is dominated by opacity shifts, with K3 line-core best representing the velocity of the spicules and with K2 features being either suppressed by the Doppler-shifted K3, or enhanced via increased radiation from the lower layers, thereby leading to a strongly enhanced K2 peak at a wavelength of opposite Doppler-shift to that of the line core. We also observe Mg II k 2796 A spicule spectra from the IRIS satellite at moderate velocities with very similar spectral properties and, using our interpretation of spicular chromospheric line-formation, produce synthetic Mg II k 2796 A profiles of spicules using the RH1.5D radiative transfer code. With unprecedented spatial and spectral resolution coupled with high cadence, the observations by CHROMIS can serve as an ideal diagnostic to explore the physical properties of spicules in great detail, thereby providing much-needed constraints for further numerical modeling efforts.