A manuscript draft should not be treated as an automatic scientific conclusion. The safe workflow is: define the research question, profile the data, choose methods with rationale, produce reviewable outputs, reconcile numbers, verify references, and let the researcher approve the final interpretation before submission.
The workflow
| Step | What happens | Review point |
|---|---|---|
| 1. Intake | Collect study design, sample size, variables, outcome, target journal range, and available files. | Confirm that the requester has rights to use the data and that sensitive data handling is in scope. |
| 2. Data profiling | Check missingness, variable types, distributions, outliers, event counts, and basic feasibility. | Flag data gaps that may change the analysis plan or make the project unsuitable. |
| 3. Analysis plan | Select candidate statistical methods and document assumptions, alternatives, and exclusion rules. | Researcher or statistical reviewer approves the plan before relying on results. |
| 4. Tables and figures | Prepare Table 1, model outputs, survival curves, forest plots, subgroup outputs, or other agreed outputs. | Numbers are reconciled across text, tables, and figures. |
| 5. Manuscript support | Draft methods/results language and provide a structured manuscript starting point. | Researcher revises wording, interpretation, limitations, and claims. |
| 6. Reference and package check | Prepare citation list, DOI/PubMed checks where available, cover letter or checklist support when in scope. | Researcher verifies final citations and journal requirements. |
What files are useful
Data files
CSV, XLSX, SPSS, REDCap-style exports, questionnaire exports, or other structured datasets. During free assessment, metadata is usually enough; do not send identifiable patient data.
Study context
Research question, primary outcome, inclusion/exclusion criteria, variable dictionary, target journal range, and any prior protocol or ethics documentation.
Existing outputs
Prior tables, draft text, figures, reviewer comments, statistical notes, or analysis code can help avoid repeating work and preserve researcher intent.
Compliance notes
Institutional restrictions, data-use agreements, de-identification status, authorship expectations, and required reporting checklists should be disclosed early.
Typical deliverables
- Data quality summary and analysis feasibility notes.
- Statistical tables and figure outputs for review.
- Methods and results draft text with documented assumptions.
- Reference list with DOI, PubMed, or journal links where available.
- Reproducibility notes or analysis code when included in the written scope.
- Submission package items such as cover letter or reporting checklist support when agreed.
FAQ
Can I use this if I only have CSV or Excel data?
Often yes, if the dataset contains enough variable definitions, outcomes, and sample information. The first step is a feasibility assessment.
Who decides the final research conclusion?
The researcher does. data2paper.ai can support analysis and drafting, but final scientific interpretation and submission decisions remain with the researcher.
Does data2paper.ai guarantee publication?
No. We do not guarantee acceptance, impact factor, review outcome, or publication speed.
Start with a metadata assessment
Send sample size, study design, variables, outcome, and target journal range. Do not send identifiable raw data during the free assessment stage.
Request assessment Download questionnaire