Dataset registry
Track identity, versions and custodians.
Data & Findings
Keep every finding connected to how it was produced.
Register datasets, versions and analyses, then capture findings with confidence, limitations and links to the tables, figures and evidence that explain them.
Open SciCollabThe problem
A result without provenance is difficult to trust or reproduce.
Files called final, final-2 and final-revised do not explain which dataset, transformation or analysis produced the number that reached the manuscript.
- Unclear dataset versions
- Tables detached from analyses
- Findings lose caveats in writing
A working sequence
Preserve the path from data to claim.
-
01
Register
Name datasets, versions and ownership.
-
02
Describe
Record transformations and analysis intent.
-
03
Capture
Turn outputs into structured findings.
-
04
Trace
Link findings to evidence and manuscript sections.
Inside the workspace
A research record around the analysis—not a replacement for it.
Analysis records
Document purpose, inputs and method.
Tables and figures
Keep outputs with their producing context.
Structured findings
Record interpretation, confidence and limits.
Provenance links
Follow a finding back to data and analysis.
Manuscript connection
Carry findings into writing with context.
Product evidence
A visible lineage from dataset to Results.
01Dataset v3
02Analysis A-12
03Finding 04 · moderate confidence
04Figure 2
05Results §2
SciCollab
Make the result traceable before it becomes a sentence.
Open SciCollab and preserve the lineage of the next finding.