Energy-efficient operation is mandatory in wireless sensor networks due to the limited energy budget of sensor nodes. Considering the potential application of wireless sensor networks in logistics, cost efficiency is another major requirement due to high cost pressure. To save on data transmissions within such sensor deployments, which account for the major part of energy consumption and monetary costs, we develop a method for data filtering based on an in-network determination of transmission relevance of sensor data. Our approach explicitly incorporates interdependencies between wireless sensor nodes and their measurements and data transmissions. It contributes to efficiency in wireless sensor networks by filtering out irrelevant data and enables a subsequent reduction of unnecessary transmissions from a network-wide view, while being able to still offer real-time data provision with sufficient data fidelity to stakeholders. The benefits of our approach are indicated by preliminary evaluation results.