Page 347 - CyberSecurityEssentialsEnglish

TABLE 7-36: THE STRUCTURE OF AN ORGANIZED RESEARCH FOLDER

Folder or file What it holds

Readme The purpose of the project, how to run it, and who owns it.

The question, the criteria, the plan, and the recorded
Protocol
amendments.

Raw data Untouched original files, kept read-only.

Processed data Outputs of cleaning, with the transformation script.

Code Executable scripts with versions and dependencies.

Evidence matrix The extraction table with source identifiers and locations.

References The verified list with the reference passports.

AI log The record of tools, purposes, outputs, and verification.

Decision log What changed, why, when, and who approved it.

Outputs Drafts, figures, and final versions, numbered by version.

TABLE 7-37: WHAT MAKES A STEP RE-EXECUTABLE

Element The question

Inputs Are the exact source files identified and preserved?

Is the step written as code or a documented instruction rather than a
Procedure
manual edit?

Environment Are the software, library versions, and seeds recorded?

Is the result saved with a name that ties it to the step that produced
Outputs
it?

Decisions Is any judgment call written down with its reason?

✕ Common misconception Correction

Leaving important steps inside A step that cannot be re-run is a step

conversations or in undocumented nobody can check. Move it into code, a
manual edits. log, or a written decision.

+ For the advanced reader: environments and automated checks

When building more advanced analysis pipelines, containers, environment files,
executable report templates, and automated data tests can be used. But reproducibility

346