技能 数据科学 论文数据与代码存档规范

论文数据与代码存档规范

v20260724
gcb-reporting-and-data-policy
本指南详细介绍了《全球变化生物学》期刊对数据和代码的强制性开放共享要求。它强调了初级和次级数据,以及所有支持性的代码/软件,必须在投稿前存入公开的、可获得DOI的公共存储库(如Dryad、Zenodo)。核心要求是:必须公开存档,不能以“可应要求提供”为由豁免。
获取技能
60 次下载
概览

Reporting & Data Policy (gcb-reporting-and-data-policy)

GCB treats open data and code as a condition of publication, not a courtesy. Primary and secondary data supporting the results must be archived in a publicly accessible repository with a persistent identifier (DOI), code/software likewise (e.g., Zenodo), and the manuscript must carry a data availability statement. Crucially, "available on request" is not accepted. Build the deposit as you go. Confirm current wording on the policy page before submission.

When to trigger

  • Writing the data availability statement
  • Choosing repositories and minting DOIs for data and code
  • Handling data that cannot be fully shared (sensitive species locations, third-party/licensed data)
  • Final reporting checks before submission

What GCB requires (verify current wording)

  1. Archive data with a DOI. Deposit primary and secondary data in a public, DOI-minting repository (e.g., Dryad, Zenodo, PANGAEA) with metadata sufficient for a third party to interpret the data correctly — before acceptance/publication.
  2. Archive code with a DOI. Code, software, and documentation supporting the results go to an appropriate public repository (e.g., Zenodo via a GitHub release) with a persistent identifier.
  3. Data availability statement. State exactly where the data and code live and how to access them; "available on request" is not sufficient.
  4. Reviewer access. Make data accessible to peer reviewers on request during evaluation.
  5. Reporting completeness. Report sample sizes, replication, units, methods, and software versions well enough to reproduce every result.

When data cannot be fully shared

  • Sensitive data (e.g., precise locations of threatened species, human-subjects or provider-licensed data): explain the restriction, give a clear access pathway (provider, application process), and share what can be shared (de-sensitized/aggregated layers) plus full code.
  • Document why the restriction applies; do not use sensitivity as a blanket reason to skip deposit.

Build-as-you-go checklist

  • Data archived in a DOI-minting public repository with interpretable metadata
  • Code/software archived (Zenodo/GitHub release) with a DOI
  • Data availability statement names repository + access (not "on request")
  • Sample sizes, replication, units, software versions reported
  • Sensitive data: restriction explained + access pathway + shareable subset
  • Manuscript exhibit numbers match the archived outputs

Repository fit by data type

Different global-change data types land best in different DOI-minting archives. Treat this as a routing guide, then confirm the current accepted list against the journal's author guidelines.

Data type Typical archive Note
Ecological tabular / experimental Dryad Curated, ecology-oriented
Code + figure pipeline Zenodo via a GitHub release Versioned, DOI per release
Oceanographic / Earth-system PANGAEA Geo/environmental specialist
Sequences INSDC (GenBank/ENA) Domain-mandated, then cite accession
Sensitive species locations Restricted deposit + access pathway Share de-sensitized layer + full code

Worked micro-example (illustrative)

A remote-sensing carbon-flux paper archives three things, not one: the gap-filled flux table to a DOI-minting repository; the processing and modelling code to Zenodo via a tagged GitHub release; and the raw tower coordinates with a stated restriction because one site is on a protected reserve. The data availability statement names each DOI and the access route for the restricted coordinates. A weak version deposits only the figures' CSV and writes "code available on request" — which GCB does not accept. The DOIs here are illustrative placeholders; mint real ones before submission.

Compliance pushback patterns and the fix

  • "Statement says available on request" → replace with named repository, DOI, and access route; GCB does not accept request-only.
  • "Code not archived, only data" → deposit the analysis/modelling code so every figure reproduces.
  • "Metadata insufficient" → add units, sampling design, and variable definitions a third party can read.
  • "Sensitive locations withheld with no pathway" → explain the restriction, give the application route, and share an aggregated layer plus full code.

Anti-patterns

  • "Data available on request" (explicitly rejected by GCB)
  • A personal website or transient cloud link instead of a DOI-minting repository
  • Archiving data but not the code that produced the figures
  • Metadata too thin for a third party to interpret the data
  • Treating deposit as a post-acceptance afterthought

Output format

【Data archived】DOI-minting repo + metadata? [Y/N]
【Code archived】Zenodo/release with DOI? [Y/N]
【Availability statement】names repo + access (not "on request")? [Y/N]
【Sensitive data】restriction explained + access path + shareable subset?
【Reproducible reporting】n, units, versions complete? [Y/N]
【Next】gcb-writing-style

Supplementary resources

信息
Category 数据科学
Name gcb-reporting-and-data-policy
版本 v20260724
大小 5.75KB
更新时间 2026-07-28
语言