The management and analysis of large-scale datasets - described by the term Big Data - involves the three classic dimensions volume, velocity and variety. While the former two are well supported by a variety of software components, the variety dimension is still rather neglected but crucial for the integration and analysis of heterogeneous scientific data. In this talk we discuss the principles of Linked Research Data and practical approaches for their implementation using semantic technologies such as the Data Spaces or Knowledge Graphs. As a concrete example, we discuss the collaborative creation of the BigDataEurope Platform, an open software stack for researchers addressing Europe's societal challenges.