With the development of Internet, the availability of various types of data (multimedia, data from databases, etc.) has increased. These data present different forms and semantics (we name them "complex data") and may be unstructured, structured, or semi-structured. In order to prepare them for proper analysis, a multidimensional modeling layer is needed. To be efficient in terms of quality of service, analysis tools and their input data should be properly structured and prepared in a previous step. Furthermore, the classical warehousing approach is not very adequate while dealing with complex data. In order to address these issues, this paper presents a generic UML model capable of representing a wide range of complex data. On the other hand, we use data mining techniques to prepare complex data for multidimensional modeling.