Maintaining data in large repositories requires a lot of effort. Type schemas change over time due to changing requirements or mistakes in the initial abstraction process. When these changes happen, objects of that type need to be updated as well to ensure they conform to the new schema. Otherwise, a program using the data would have to treat every object differently, depending on the schema version. The transformation of all objects to conform to a new schema is called data migration. MUSE (Muster Suchen und Erkennen, eng.: Search and Recognize Patterns) and MUSE4Music are specific repositories for costumes and classical music. These repositories are used to prepare categorical data for automatic analyses. MUSE4Anything, successor of these implementations, is a generic data repository that allows managing domain specific data conforming to user-defined ontologies. In this environment, defined types and their requirements may also change over time. While type and object modifications are already supported in MUSE4Anything, there is no implementation of automatic migration of objects after a type update yet. This thesis aims to extend the MUSE4Anything system with a migration engine that performs these processes. By automating data migration, the required manual effort is significantly reduced, and the user experience, especially for huge repositories, is improved.