RESEARCHER GUIDE & DOCUMENTATION

from your data to your verified report, step by step.

Analysez AI helps researchers inspect data quality, review every cleaning decision with explicit consent, select defensible statistical tests, and produce verified reports for academic and clinical studies.

Start a study →Connect your AI tool (Claude / ChatGPT)

01 · GETTING STARTED

Create a study and set your research objectives

Start by entering your study title and objectives or uploading your research proposal (Word or text PDF), then upload your dataset (CSV, TSV, Excel, SPSS, or Stata). The AI guide organizes your study objectives and variable roles before any numbers are run.

1. Define the objective: Review and confirm your numbered research objectives and study design.

2. Upload your dataset: Select your data file (or specific workbook sheet) to begin the data quality check.

3. Follow the five guided steps: Objectives → Data cleaning → Descriptive statistics → Inferential tests → Verified report.

02 · PREPARING YOUR DATA

Inspect your data quality; nothing changes without your consent

The cleaning review highlights missing values, out-of-range entries, mixed Arabic/English category labels, duplicates, and personal identifiers. Every suggested fix stays a draft until you inspect the cleaned-table check and explicitly confirm saving a new dataset version.

Supported file size limits

  • • File size: up to 250 MB
  • • Rows: up to 1,000,000 rows
  • • Total cells: up to 30,000,000 cells
  • Tip: Please split very large files that exceed these limits into smaller subsets before uploading.

03 · ANALYSIS AND REPORTS

Generate verified tables, statistical tests, and publication reports

Our statistics engine computes baseline descriptive statistics (Table 1), checks distributional assumptions to select the right hypothesis test or regression model, and prepares an APA-formatted report with a verifiable certificate and downloadable files.

Descriptive (Table 1)

Mean ± SD for normal variables, Median (IQR) for skewed variables, and counts with percentages for categories.

Hypothesis tests & models

Assumption checks before running tests, returning exact p-values, 95% confidence intervals, and effect sizes.

Reports & certificates

Download your formatted report, cleaned dataset, figures, and verification certificate with a Study Pass.

04 · CONNECT YOUR AI TOOL

Connect Claude, ChatGPT/Codex, or Cursor for read-only verified biostatistics

The official Analysez AI connector lets your AI assistant run verified biostatistics on your studies or CSV snippets. Every tool is strictly read-only — it never alters or deletes your files, and personal identifiers are redacted automatically before results are returned.

Claude (Web, Desktop & Claude Code)ChatGPT & OpenAI CodexCursor

Official Connector URL & How to Connect

https://app.analysez.com/api/mcp
  1. Sign in: Add the connector in Claude or ChatGPT using https://app.analysez.com/api/mcp, or open AI Connectors in your Account Settings.
  2. Approve connection: Review the read-only permissions and redirect hostname on the consent screen, confirm your password, and click Approve & Connect.
  3. Ask your AI tool: Ask your assistant to check your study data quality, build Table 1, run statistical tests, or audit your manuscript numbers.

What you can do with the connector (all read-only)

Browse your studies

Free

View the studies in your account, their current step, and whether a cleaned dataset is ready.

Example: "Show my studies on Analysez AI and tell me which ones have cleaned data ready."

Check your data quality

Free

Run a thorough, read-only quality check to spot missing values, unusual entries, coded missing values, mixed Arabic/English labels, duplicate rows, and personal identifiers without changing a single cell.

Example: "Check the data quality of my uploaded diabetes study and summarize any issues needing attention."

Pick the right statistical test

Free

Evaluate variable types, sample sizes, distribution shape, and group structure to recommend a defensible statistical test before running analysis.

Example: "Which statistical test should I use to compare hba1c_6m across study_arm?"

Verify summary numbers

Free

Recalculate p-values, test statistics, odds ratios, and integer mean consistency from published summary numbers without needing raw participant rows.

Example: "Verify whether Group 1 (mean 21.0, SD 2.37, n 6) vs Group 2 (mean 7.0, SD 1.41, n 6) matches a reported p-value of 0.45."

Build Table 1 (baseline characteristics)

Paid plans

Generate a publication-ready descriptive summary table with Mean ± SD, Median (IQR), and category counts and percentages, optionally stratified by study group.

Example: "Build a baseline Table 1 stratified by treatment_group for age, bmi, and smoking_status."

Run statistical tests and models

Paid plans

Run verified group comparisons, paired tests, correlations, and regression models with exact test statistics, two-sided p-values, 95% confidence intervals, and effect sizes.

Example: "Run the verified hypothesis test comparing systolic_bp across study_arm and report the effect size and 95% CI."

Check draft manuscript numbers

Paid plans

Cross-check sample sizes (N), averages, percentages, and p-values written in your draft Results section against the numbers computed from your dataset.

Example: "Check this draft paragraph against my study data: In N = 120 patients, mean age was 54.2 and p = 0.012."

Prepare methodology justification and checklist

Paid plans

Produce a clear justification for why the statistical test was chosen, compare parametric and non-parametric alternatives, estimate detectable effect size at 80% power, and prepare a CONSORT or STROBE checklist.

Example: "Prepare a methodology justification and STROBE checklist for comparing length_of_stay across exposure_group."

How to disconnect or revoke

Open Account Settings → AI Connectors and click Revoke next to any active connection to invalidate its token immediately, or disconnect Analysez AI inside your Claude or ChatGPT settings.

Privacy & per-user access

Every connection is scoped strictly to your own account and is read-only. Direct personal identifiers (national IDs, phone numbers, emails, names) are automatically redacted, and your data is never used to train AI models.

Troubleshooting & support

If a connection expires or returns an authorization error, reconnect or generate a fresh key in Account Settings. For datasets over 5 MB, upload directly on the web workspace. Need help? Email info@analysez.com.

05 · PRIVACY AND SECURITY SUMMARY

Your research data stays yours and is never used for AI training

All data is encrypted in transit and stored in isolated, access-controlled storage. We never sell research data or use customer datasets to train AI models. You can delete any study or revoke external connections at any time.

Read the Privacy Policy →Read the Terms of Service →

06 · FREQUENTLY ASKED QUESTIONS

Answers to common questions

Does the AI guess statistical numbers or change my data automatically?

No. Every statistical result and p-value is computed by our verified statistics engine, and no cell in your dataset is ever modified without your explicit review and confirmation.

What is included in the free tier vs. a paid Study Pass?

The free tier includes study browsing, data quality checks, statistical test selection, and summary-number verification. A Study Pass (Project Pass, Thesis, or Research Lab) unlocks interactive deep cleaning, Table 1 generation, full hypothesis and regression tests, manuscript auditing, and report downloads.

Do follow-up questions on the same dataset through the AI connector use extra study credits?

No. All follow-up analyses on the same dataset belong to one study and never consume additional study credits.