ABSTRACT: The Bicycle Sharing Systems (BSS) o er a mobility service in which public bicycles are available for shared use. The demand of this type of systems is characterized to be unpredictable, asymmetric and spatial-time dependant. These demand characteristics a ect the system balance during specic periods of time. That is, bicycles accumulate in some stations, leaving no free parking docks for incoming users, while other stations are empty not being able to satisfy new users demand. The repositioning of bicycles is the most used strategy to balance the system. In that strategy, external vehicles transport bicycles from crowded stations to empty stations in which a demand peak is foreseen. Usually, the operational área is divided into zones to be served by dierent repositioning vehicles. This paper addresses the districting problem arisen when creating the repositionig zones. It deals with a tactical decision (i.e., districting problem), unlike most of the research on repositioning bicycles in BSS, which focuses on operational decisions such as routing and inventory management. Moreover, this work main contribution is to take into account when defining the districts of the BSS, not only distance and connectivity, but also criteria such as demand patterns and stations criticality. A mathematical model that involves those criteria is proposed. It is tested on instances built from real operational data of ECOBICI in Mexico City, which allows to draw insights to be taken into account by the system operators.COL0031851