Flow kinematics of granular materials in a hopper/ silo is complex and strongly influenced by intrinsic particle features such as size, shape, and interparticle friction. The current study focuses on understanding the influence of particle morphology, interparticle friction and hopper exit width on spatial and temporal variations in velocity. Using non-convex polygons digitised from high-resolution images, the flow characteristics, and velocity fluctuations are explored inside a conical hopper. Coarse graining was employed to obtain continuum fields from discrete quantities. K-means clustering (KMC), an unsupervised algorithm, is used to cluster the spatial variation of coarse grained velocity inside the hopper. The three flow regimes classified using KMC consist of the bottom-most cluster being in the shape of a semi-circular arch (Free Flow Cluster - FFC), below which particles loosely interact and move at higher velocities than the rest of the hopper. The intermediate cluster describes the development of funnel flow, while the top cluster exhibits mass flow where the particles flow at lower velocities. The morphology and friction of the particles affect the shape and size of these three regimes. In all hopper widths, the FFC height was lowest for bulky-shaped particles because of their high sphericity and roundness values. The coarse grained velocity was later used to characterize the temporal fluctuations in the hopper. The velocity waves were observed to propagate from the hopper exit to the bin section. The wave was observed to have clear periodicity in the bin section of the hopper and also strongly influenced by particle shapes. Further, the discharge rate, coarse grained orientation, angular velocity and stress were also compared for all the morphologies during the flow.