Reproduction
Can the same approved inputs, code and environment reproduce the output?
Analysis / Approved recurring routines
Repeat an approved
analysis
Schedule an already approved analysis, capture its inputs and environment, and prepare diagnostics, figures and tables for a scientist to review.
Walk through the example ↓An illustrative scenario
A field team repeats the same agreed summary of instrument measurements every week
01 / Supplied by your team
An approved method and a versioned input
- Analysis specification
- The scientist defines the question, estimand, method and expected outputs.
- Approved routine
- A reviewed script, its parameters and a recorded software environment.
- Input snapshot
- DEMO-WEEK-04: an illustrative, versioned instrument export.
02 / Automated preparation
Run, record, and surface exceptions
- Scheduled execution
- Run the approved routine against the named input version.
- Diagnostics
- Surface missing values, failed checks and changes from the approved assumptions.
- Output bundle
- Prepare figures, tables and a manifest connecting code, parameters and input data.
Ready for the researcher to review.
03 / A human decision
The scientist decides what the result means
Sample exception
The illustrative diagnostic sheet flags a gap in one sensor’s observations. An interpretation cannot be signed off automatically.
The sample reviewer accepts the descriptive output for inspection and leaves interpretation pending until the data gap has been assessed.
This control demonstrates a sample decision. Nothing is sent or saved.
04 / A clear handoff
A reproducible review package
The sample review decision is still pending.
Sample decision applied
Release descriptive outputs for inspection; interpretation remains pending.
- Run manifest
- Named data, code, environment and parameter versions.
- Figures + tables
- Prepared outputs with units, labels and the associated diagnostics.
- Review note
- What passed review, what remains uncertain and who owns the next decision.
Apply the sample review decision to enable this download. Illustrative handoff contents. A real project needs an agreed scope and review.
Explore each stage. Decisions here only change this demonstration.
Researcher decisions
- The estimand and research question
- Statistical methods and assumptions
- Interpretation and claims
What a first pilot would measure
Test reproducibility
One recurring approved routine, replayed on a fixed dataset before any schedule is enabled.
Review burden
Researcher time spent checking a run, alongside errors found or missed.
These are proposed evaluation criteria. No client results or performance gains are claimed.
Plan an analysis workflow
Tell us how the task works today. We’ll agree the inputs, review steps, scope and price with you.