The concept of Precision Dairy Farming, as an integrative approach to dairy farming, sets high standards for data management. Special consideration during its implementation must therefore be given to support both operational tasks, as well as analytical tasks (e.g. OLAP). Hence, we propose a parallel modelling procedure in a common design process. The approach taken is a hybrid modelling process integrating operational and analysis-oriented data. This results in a joint relational data scheme as a basis for operational and analysis-oriented information systems. Starting point is an analysis of the information requirements in order to determine the requirements of the information system. The implemented analysis methodologically focuses on expert interviews in order to identify future needs. Based on these criteria, the next step is semantic data modelling. We used the Entity-Relationship Model in order to map the operational context. In the course of modelling the analysis-oriented data, the multidimensional Entity-Relationship Model was also applied, due to the lack of rich semantics to map analytical relations. However, these notational elements do not sufficiently represent the application-related analysis-oriented perspective. Therefore, we introduced new notation elements. The semantic modelling is followed by a relational modelling. Based on separate data schemes on a semantic level, relational data schemes were created. Due to the integration of both relational data schemes, we achieved a common data schema. The result of this hybrid modelling process is a relational data scheme which describes both operational, as well as analysis-oriented data. However, the specific advantage of this hybrid modelling approach has to be seen in the close interconnection of both data views on a data level. The integration within the design process reduces the risk of redundancy and inconsistency caused by conflicts of schemes in the course of the implementation of the operational and analysis-oriented databases. Furthermore, the conditions of operational consistency are transferred to an analysis-oriented data view. This is an important step to increase the quality of analysis-oriented databases. This paper describes the methodology of hybrid data modelling and beyond that, presents a reference model in the area of dairy farming. |