Zeolites form a class of microporous aluminosilicates of great interest due to their multifarious applications in industry and everyday life. Their porous structure allows small molecules to be adsorbed and to diffuse inside crystals, and depending on the zeolite type and on the diffusant species a variety of behaviours is possible. Molecular Dynamics is now widely used in order to understand the microscopic mechanisms of adsorption and diffusion occurring within these materials as well as in MOFs and ZIFs. A major drawback of MD for this kind of systems is its high computational cost, so that coarse-grained methods, speeding up simulations without losing the essential features of dynamics, are valuable tools for exploring the behaviour of guest molecules on time and space scales hardly, if at all, reachable with ordinary MD.<br/> The first step in our proposed method is the clustering of MD trajectories to obtain a discretized version of the motion of adsorbed molecules within the zeolite. Each pore in the aluminosilicate is partitioned in a number of regions and each point in the original trajectory is mapped to the proper region based on a distance criterion. The regions correspond roughly to the main basins in the potential energy surface (PES).