Data Science Orchestrator: A Containerized Trusted Research Environment for Flexible and Secure Analytical Pipelines

Area of Science:

• Data Science
• Computational Research
• Research Infrastructure

Background:

• Researchers often use personal laptops with non-versioned datasets and unvetted software, risking data leaks and hindering result reproducibility.
• Current practices for data analysis are often slow, tedious, and lack robust security and version control.

Purpose of the Study:

• To introduce the Data Science Orchestrator (DSO) as a solution for secure and reproducible data analysis.
• To expand the DSO's capabilities to address broader research infrastructure needs beyond initial data analysis.

Main Methods:

• Development of the Data Science Orchestrator (DSO) as a turnkey solution for deploying analysis tools.
• Analysis of researcher requirements to identify needs for federated analysis and automated reporting pipelines.
• Implementation of automated analysis pipelines for repeated reporting at different intervals.

Main Results:

• The DSO provides a secure and reproducible environment for data analysis.
• The DSO's scope has been expanded to include federated data analysis capabilities.
• Automated analysis pipelines enable repeated reporting and extend the DSO's impact.

Conclusions:

• The Data Science Orchestrator significantly improves research safety, reproducibility, and efficiency.
• The expanded DSO addresses critical research infrastructure needs, supporting advanced data analysis and reporting.