Machine-readable datasets that have increasingly become available in open formats in recent years have great potential as a foundation for innovative applications and services. Linked Data in particular - a set of best practices for publishing and connecting structured data on the Web - has facilitated significant progress in evolving the Web of documents into a Web of Data. However, although this concept has opened up many opportunities for data sharing and collaboration, integrating data is still a challenging task that requires considerable technical expertise and a profound understanding of the underlying datasets. In this paper, we introduce a novel approach to provide knowledge workers with the necessary tools to leverage the fast growing Linked Data Cloud by creating semantic-aware dataflow processes. To this end, we introduce the"Linked Widget"concept as an enhancement of standard Web Widgets. These widgets are based on a semantic data model that facilitates powerful mechanisms for gathering, processing, integration, and visualization of data in a user-friendly Mashup environment. By allowing knowledge workers to easily create complex Linked Data applications in an ad-hoc manner, our approach should contribute towards reducing existing barriers of Linked Data adoption.