Structuring data analysis projects in the Open Science era with Kerblam!

Area of Science:

• Data Science
• Computational Biology
• Research Reproducibility

Background:

• Data analysis project structuring often relies on personal preferences, hindering accessibility and transparency.
• Open Science principles advocate for more open, understandable, and reproducible research practices.
• Applying Open Science to project structure can improve research accessibility and collaboration.

Purpose of the Study:

• To investigate common data analysis project structures and identify best practices.
• To develop fundamental design principles for structuring data analysis projects.
• To introduce Kerblam!, a tool to support these principles and improve data management.

Main Methods:

• Analysis of data analysis project templates from GitHub repositories.
• Visualization of consensus structures to identify common patterns and characteristics.
• Development of design principles based on Open Science philosophy.

Main Results:

• Limited overlap exists among project templates, but distinct practices were identified.
• Fundamental design principles for structuring data analysis projects were formulated.
• Kerblam! was developed as a project management tool to facilitate data handling and workflow execution.

Conclusions:

• Adherence to proposed design principles and use of Kerblam! can increase transparency and understandability of data analysis projects.
• The proposed framework aims to make data analysis projects more useful to the wider research community.
• Promoting standardized project structures fosters reproducibility and facilitates collaboration in data science.