Scalable Infrastructure for Validation of Computational Social Science Research (SIVACOR)

Scalable Infrastructure for Validation of Computational Social Science Research (SIVACOR) aims to build a scalable framework that enables users to confirm and demonstrate reproducible analysis. We build on the metadata schema developed within the TRACE project to build a system that generates a Trusted Research Object (TRO). Researchers will use the public-facing SIVACOR system to execute their analysis, generating a certified record of successful execution. The certificate assures researchers that their models are likely to run when executed by others, and can use the generated certificate to demonstrate this fact when submitting articles to a journal. Our target audience are social science researchers who ultimately need to publish in a journal that has a data and code verification process in place, but generally do not have the technical expertise to work in containerized environments or larger computational pipelines.

SIVACOR allows researchers to work in their preferred environments until the last moment, without needing to fully understand Docker or similar technologies (which remain little used in the social sciences). The system will provide a web interface where users can upload a ZIP file containing their code and data, and select from a set of curated computational environments. The SIVACOR backend will then automatically create a containerized environment, execute the analysis, and generate a TRO that includes the code, data, execution environment details, and results. This TRO can then be accessed by journal editors, replacing the often tedious and manual process of verifying the reproducibility of the submitted research.