FAIR data pipeline: provenance-driven data management for traceable scientific workflows

Area of Science:

• Epidemiology
• Data Science
• Scientific Computing

Background:

• Epidemiological studies rely on data, but managing rapidly changing and imprecisely identified data is challenging.
• Lack of transparency in data provenance erodes public trust in science-based policy decisions.

Purpose of the Study:

• To demonstrate a Findable, Accessible, Interoperable, and Reusable (FAIR) data pipeline.
• To enhance transparency and traceability in scientific research, particularly for policy-facing applications.

Main Methods:

• Development of a FAIR data pipeline for annotating data during analysis.
• Implementation of a system to trace scientific outputs back to primary data through source code.

Main Results:

• The pipeline facilitates easy annotation of data as it is used in analyses.
• It enables tracing the provenance of scientific outputs back to the original data and analytical code.

Conclusions:

• The FAIR data pipeline enhances transparency by clarifying the evidence supporting scientific claims.
• This tool empowers scientists and the public to better assess scientific evidence and supports policymakers in justifying decisions.