A Semantic Workflow Approach to Web Science Analytics.

Reproducibility and reuse are rapidly becoming guiding principles in publishing and sharing scientific results. In order to enhance researchers' ability to leverage existing results, many are moving in the direction of semantic workflow systems, which enable users to define and share experimental procedures as linked data on the Web. These workflows provide a powerful mechanism for reproducing experiments and thus are well-suited for Web Science tasks. In order to aid users in the process of experiment design and reproduction, we are integrating the Workflow INstance Generation and Specialization (WINGS) system with our existing Semantic Numeric Exploration Technology (SemNExT) framework. This will provide a completely open-source stack for designing in silico experiments using a combination of semantic and numeric analyses. We will explore how this system may be configured to create reproducible Web Science workflows, especially as it pertains to data federation across the Web. We are leveraging our existing tooling as we develop new approaches for automatically generating provenance for interacting with remote endpoints of heterogeneous data sources. This will support not only the aggregation of diverse and geographically-disparate data sources across the Web, but also collaborative science by allowing other users to reproduce and expand on the same results using shared workflows.