Skip to content
MathJet Docs

Building a reproducible analysis end-to-end

40 minutes · Capstone tutorial. Assumes Tutorials 1, 2, and at least one of 3, 4, or 5.

The earlier tutorials each demonstrated one of MathJet®‘s differentiators in isolation. This is the one where they combine. You’ll build a complete reproducible analysis from raw data to a presentable result in a single .mjw workspace — loading data from an Excel file, preprocessing in Python, modeling in R, and visualizing the result with axis folding and the graph companion table. Along the way you’ll use action recording to capture GUI manipulations as reproducible code, save the workspace, reopen it, and verify that every piece of state — variable values, plot zoom, cell selection — is restored exactly. The tutorial closes with an honest assessment of what “reproducible” means in MathJet’s model today, where it’s strong, and where the workflow-fusion layer is still being expanded with each release.

What you’ll learn

Section titled “What you’ll learn”
  • How to structure a multi-step analysis as a single .mjw workspace that holds data, code, plots, and state together
  • How to combine data loading (.xlsx or CSV), Python preprocessing, R modeling, and interactive visualization in one session — without restarts or serialization
  • How to use action recording to capture GUI manipulations as reproducible Python or R code, and how to replay them
  • How to save a workspace, share it, and reopen it with all variable values, plot states, and selections restored
  • What “reproducible” means in MathJet’s model — the boundary between what’s captured automatically, what requires action recording, and what isn’t yet capturable today

What you’ll need

Section titled “What you’ll need”
  • MathJet installed (download)
  • Tutorials 1 and 2 completed, plus at least one of Tutorial 3 (MATLAB), Tutorial 4 (Jupyter), or Tutorial 5 (R / RStudio) — this capstone assumes you’ve used the basic features and at least one language layer before
  • A realistic dataset from your own work, or the provided measurements.csv extended with grouping variables for richer modeling
  • An hour set aside — the walkthrough is designed to be done in a single sitting, and the value is in seeing the pieces compose, not in stopping halfway

This tutorial is in development and will be published soon. In the meantime, see the Features page for an overview of the capabilities this tutorial covers.