NoSQL data stores are becoming widely used to handle Big Data; these systems operate on schema-less data model enabling users to incorporate new data into their applications without using a predefined schema. But, there is still a need for a conceptual model to define how data will be structured in the database. In this paper, we show how to store Big Data described by conceptual model within NoSQL systems. For this, we use the Model Driven Architecture (MDA) that provides a framework for models automatic transformation. Starting from a conceptual model describing a set of complex objects, we propose transformation rules formalized with QVT to generate NoSQL physical models. To ensure efficient automatic transformation and to limit the impacts related to technical aspects of NoSQL systems, we propose a generic logical model that is compatible with the three types of NoSQL systems (column, document and graph). We provide experiments of our approach using a case study related to the hea lth care field. The results of our experiments show that the proposed logical model can be effectively transformed into different NoSQL physical models independently of their specific details.