Data Strategy

This document outlines the standardized data and code management practices for research and advisory projects within our computational limnology team.

📁 Project File Structure

Each project must follow a consistent folder and documentation structure similar to this:

├── README.md             # Project overview and objectives
├── data/                 # Organized into raw/, processed/, external/
│   └── README.md         # Data descriptions and sources
├── scripts/              # Analysis and workflow scripts
├── outputs/              # Results: figures, tables, model output
├── reports/              # Reports, manuscripts
├── figures/              # Figures, visualizations
├── targets.R             # Workflow pipeline (e.g., {targets} in R)
├── renv/ or environment.yml # Environment/dependency management
└── devlog.md             # Log of decisions and progress (optional)

🔄 Version Control & Backup

• All projects must have a GitHub repository:

  • Public for research projects

  • Private for advisory/confidential projects

• Repositories must be mirrored locally on AU’s O:\ drive for backup

• (Ideally) Use standardized naming (e.g., project-<year>-<topic>)

🧪 Reproducibility Standards

• Use {targets} (R) or snakemake (Python/bash) for all analysis pipelines

• Final results must be fully reproducible from raw data

• Use renv, virtualenv, or pipenv for dependency management

• Clearly comment all scripts and functions

📦 Data Publishing & Archiving

• On completion, archive the GitHub repository to Zenodo to mint a DOI

• Include a CITATION.cff file for machine-readable citation metadata

• Tag final release version for reproducibility and reference

✅ FAIR Data Principles

All datasets should follow FAIR guidelines:

• Findable: Documented, discoverable via DOI

• Accessible: Public repositories or open-access archives

• Interoperable: Use open formats (CSV, NetCDF); standard units

• Reusable: Include metadata (metadata.yaml or data_dictionary.csv)

Each dataset must be accompanied by meta-data:

• Description of variables (names, units, meaning)

• Sampling methods and instruments

• Collection date, location, and processing steps

🧾 Documentation & Communication

All code must be well-commented and follow team style guides

• Preferably using {roxygen2} for commenting

Project documentation includes:

• README.md – Overview

• devlog.md – Key decisions and changes

• GitHub Projects – Plan and track decisions and work packages

• GitHub Issues – Task and discussion tracking (encouraged)

⚙️ Automation & Tools

Automate repetitive tasks where possible:

• Convert often used scripts or functions into packages/libraries (R, Python)

• GitHub repo creation, license, and structure setup

• Sync to O:\ drive using post-commit hooks or scheduled scripts

• DOI publishing via GitHub-Zenodo integration