Module 10 — Reproducibility

Question: six months from now, can you — or a reviewer — reproduce every number in your report? In ChakataStat the answer is built in: every module of this course was quietly writing a script.

The Syntax console

Open the Syntax tab. Everything you ran through the menus is there as commands — the same ones each module's "syntax trail" showed. This console is editable and executable:

1. Click Run all: the entire session replays, top to bottom, and the output regenerates.
2. Edit one line — change a variable, add one — select it, run just that.
3. Save the script as a .cks file next to your dataset.

Dialogs are how you find an analysis; syntax is how you keep it.

The reproducible bundle

A finished piece of work is three small files:

• study.ckd — data plus every variable definition (modules 1–2);
• study.cks — the ordered commands that produce the results;
• the exported output (File → Export Output…, HTML/Markdown/LaTeX) — what the numbers looked like when you reported them.

Anyone with the first two can regenerate the third — that is the standard your future self will thank you for. (Batch users can even run it headless: chakatastat --run study.cks study.ckd --out report.html.)

Capstone exercise

Reproduce a mini-report start to finish, in syntax:

1. Open the sample. In the Syntax tab, write (or paste from your trail):

   descriptives variables=[Age, BMI, SystolicBP]
   crosstabs row=SmokingStatus, column=Hypertensive, chi_square=true
   anova variables=[SystolicBP], factor=SmokingStatus, posthoc=true
   linear_regression dependent=SystolicBP, predictors=[Age, BMI]
   histogram variable=BMI
   

2. Run all, review the output, export it to HTML.

3. Save the .cks and the .ckd, close the app, reopen, run the script again — confirm the report reproduces exactly.

Where next

• Analyzing Data — the full procedure catalog, from nonparametrics to survival and time series.
• The AI assistant — plain-language analysis over the same engine, with the same syntax trail.
• Reproducibility in depth — batch runs, Python export, output formats.

You now know the whole loop: open → describe → visualize → test → transform → report — reproducibly. That is the course.