Abstract
While the number of Linked Data (LD) datasets grows, the support for their consumption is still quite limited. Data publishers expect high reuse of their LD in different applications.
At the same time, their users expect that the applications will be reusable for different LD datasets. However, the reality is different because existing LD often use proprietary vocabularies and unique combinations of non-proprietary ones.
In this paper, we introduce a backend of a platform which interconnects LD datasets on one side with applications for previewing the datasets on the other. The platform backend is automatically discovers so called application pipelines for a given set of datasets.
Each application pipeline specifies a sequence of transformation steps which transform the original data to a form required by an application. The platform is also able to execute a chosen pipeline and provide the result to the application in a prepared SPARQL endpoint.