experiment-structure

A canonical folder layout for reproducible experiments includes directories for configs, src, data, runs, analysis, and figures. This structure enforces provenance rules and provides an auditable map for where project files belong.

To scaffold a new computational experiment, generate the canonical directory tree containing configs, src, data, runs, analysis, figures, and documentation. This establishes required conventions immediately and ensures the project starts from an auditable structure.

You need a provenance structure for computational experiments when starting new work or when a process requires a destination directory. It ensures your project remains organized, auditable, and reproducible by enforcing where data, source, and analysis files belong.

Yes, this experiment scaffold works without external dependencies. It defines the required directories and provenance rules for project organization natively, ensuring your computational experiments remain reproducible without needing additional frameworks.

In a reproducible project layout, configuration files go in the configs directory and analysis files go in the analysis directory. This canonical experiment folder structure provides a single authoritative answer for where every project artifact belongs.