Data models are abstract models that standardize data formats and relationships. In other words, data models describe the concepts that belong to a certain application domain (e.g., "Device" and "Farm" are concepts that belong to the "Agriculture" domain, and "Device" is also a concept in the domain of "Smart Cities"). Over the years, many data models (and ontologies) have been produced for the precision agriculture domain. On the one hand, such models provide standards for data transmission and representation. On the other hand, these models are not suited for (automated) data integration and analysis, which are core tasks in building decision support systems for precision agriculture ---digitalized systems that support farmers and technicians in making data-driven decisions. Following the advancements in big data technologies and internet of things systems, managing such systems is increasingly harder and requires not only standards to transmit and represent the data, but also to automatically integrate heterogeneous data into a uniform medium and to automate data analysis and fruition. While this is a well-known issue in the field of precision agriculture, where data models usually fuel data silos for ad-hoc independent applications (e.g., smart watering management, autonomous weeding systems, vegetation index computation), the synergy with computer science and database techniques could both answer these challenges and open novel research directions. In this poster, we (i) describe some of the state-of-the-art models for precision agriculture and their application (e.g., from the FIWARE ecosystem), (ii) factorize the limitations and issues of such models (e.g., inter-domain ambiguities, intra-domain inconsistency, wrong modeling practices), (iii) show how computer science techniques (e.g., entity resolution, data normalization, data provenance collection) can answer these issues, and (iv) introduce novel data-driven research directions for building unifying decision support systems.