6  Tools for the Teaching Community

Beyond my own courses, I develop and maintain several technical tools aimed at supporting university instructors in statistics and data science. These open-source projects are designed to reduce technological friction for adopting reproducible practices.

6.1 Teaching Resources for Statistics and Data Science Instruction

I provide the community with a central repository that serves as a “starter kit” for building modern course websites with Quarto. This project offers a simple entry point for colleagues who want to modernize course material without needing to become web development experts.

  • Website structure and available resources:

    • Ready-to-use templates: Quarto (.qmd) templates for Reveal.js presentations, formatted assignments, and reproducible exams.
    • Technical tools: libraries of useful R functions for teaching and guides for integrating GitHub in class.
    • Step-by-step tutorials: detailed guides for getting started with R, RStudio, Quarto, and Git in a teaching context.

  • Philosophy: enable instructors to focus on pedagogical content (substance) by providing robust, turnkey technical infrastructure (form).

  • Access and documentation: https://aureliennicosiaulaval.github.io/site_ressources_SSD/

6.2 R Package UlavalSSD

To support data science teaching at Universite Laval, I created the R package UlavalSSD.

  • Problem: students (and instructors) lose valuable time managing file paths, character encoding, and multiple dependency installations.
  • Solution: a single package that distributes:
    • all pedagogical datasets used in courses (cleaned and documented),
    • helper functions for tool installation,
    • preconfigured RStudio project templates.
  • Impact: this package is now used not only in my courses, but also by colleagues teaching related courses, helping standardize the software environment.
  • Access: GitHub - UlavalSSD

6.3 Package contextR: AI Pedagogy

Currently in development, the contextR package is an experimental tool for instructors who want to integrate LLMs locally in R (see Section 4.2).

  • Objective: provide a simple interface so students can obtain a contextual interpretation of statistical output (e.g., lm regression summary) generated by AI, but strictly framed by pedagogical prompts defined by the instructor.
  • Teaching use: supports activities where students critique AI interpretations, turning the tool into a Socratic debate partner.
  • Access: GitHub - ContextR

6.4 Package tutorizeR: Automation of Interactive Tutorials

The tutorizeR package is a pedagogical productivity tool designed to quickly transform static content into interactive learning experiences.

  • Problem: creating interactive tutorials (such as learnr) often requires time-consuming rewriting of existing course material.
  • Solution: this package automatically converts existing Quarto (.qmd) or R Markdown documents into interactive modules, generating exercises and validation tests from source code.
  • Features:
    • Supports conversion to learnr (Shiny) or quarto-live (WebAssembly, serverless).
    • Transforms code blocks into exercises with solutions and automatic validation.
  • Access: GitHub - tutorizeR