Educational Innovation

Vision

I see pedagogical innovation as a lever for transforming learning.
My goal is to design environments where students become active participants in their own education through active approaches, modern digital tools, and a culture of reproducibility.

My projects are organized around three axes:

  1. Modernizing courses through open technologies (R, Quarto, Learnr, GitHub).
  2. Support and supervision through TA training and collaborative resource development.
  3. Responsible integration of artificial intelligence in teaching and student success support.

🌱 For me, pedagogical innovation means bringing together rigor, creativity, and accessibility so that every student can learn by exploring, experimenting, and understanding.

Key projects

Redesign of data science courses

Complete update of introductory data science and scientific programming courses.
Objectives:

  • make learning more active and contextualized;
  • develop francophone case studies;
  • provide automated and formative feedback through Quarto and GitHub Classroom.

Outcome: improved motivation and student participation, while promoting code reproducibility.

Improving student success support

Development of a structured training model for teaching assistants and an internal platform gathering resources, short videos, and a FAQ.
Outcome: professionalization of the TA role and greater consistency in pedagogical support.
Link: TA space - internal platform (Ulaval)

GPT-CDA project - Artificial intelligence and learning

Development of a specialized conversational assistant for support in mathematics and statistics.
Objectives:

  • provide explanations adapted to course materials;
  • detect misunderstandings and redirect to a human tutor;
  • optimize TA intervention time.

Outcome: deployment of a supervised prototype combining AI and personalized pedagogical support.

Open educational resources

I develop an ecosystem of open digital resources to support teaching and learning in data science.
This platform is designed to facilitate the creation of reproducible, interactive, and collaborative material, while promoting the dissemination of francophone resources.

1. Educational resources platform

Link: site_ressources_SSD

Created as an online library, this platform centralizes a set of resources for instructors and students.
It emphasizes reproducibility, collaboration, and sharing of teaching practices.

Content:

  • Quarto templates for websites, reports, and reproducible documents;
  • Examples of interactive tutorials built with learnr;
  • UlavalSSD R library containing functions and datasets for statistics teaching;
  • GitHub procedures and guides for version control and open sharing of teaching material.

Outcome: this platform improves coherence in teaching practices and supports collaboration among colleagues while reducing preparation time.

2. tutorizeR

Link: tutorizeR

The tutorizeR package automates conversion of Quarto (.qmd) and RMarkdown (.Rmd) files into interactive tutorials using learnr or quarto-live.
It was designed to help instructors quickly create dynamic activities without extra technical effort.

Main features:

  • Automatic conversion of documents into interactive tutorials;
  • Addition of answer zones and immediate feedback via gradethis;
  • Generation of ready-to-deploy environments (online courses, self-training, labs).

Objective: promote active and formative pedagogy by integrating interactive exercises directly into Quarto documents.
Outcome: tutorizeR substantially reduces the technical workload of tutorial creation while improving reproducibility and student engagement.

3. ZeroWasteData (in development)

Link: ZeroWasteData

ZeroWasteData is an interactive application developed in Python with Streamlit, designed to encourage smart data reuse in data science education.
It is based on a simple idea: rather than accumulating files and exhausting storage resources, we should make better use of data we already have.

How it works:

  • the user imports a dataset;
  • they indicate which analyses or activities have already been performed on that dataset;
  • the application automatically suggests additional possible analyses and links them to relevant statistical or data science concepts (for example: visualization, correlation, regression, supervised learning, clustering, etc.).

Educational objectives:

  • increase the value of existing datasets by exploring new analytical perspectives;
  • reduce duplication and digital waste, connecting pedagogy and data sustainability;
  • foster active discovery by linking each analysis type to a concrete teaching concept.

Expected outcome:

A user-friendly and open-source Streamlit application acting as an intelligent teaching assistant, helping instructors and students get the most from a single dataset,
while promoting responsible and sustainable management of digital resources.

4. Interactive learnr tutorials

Links:
- 🚴‍♂️ BIXI tutorial - Data analysis and visualization
- 🎨 ggplot2 tutorial - Data visualization

These interactive tutorials, developed with the learnr package, allow students to learn data science fundamentals by directly manipulating R code in a guided environment.
They combine explanations, practical exercises, and automatic feedback to support active learning and progressive understanding of concepts.

BIXI tutorial
This tutorial is based on real-world BIXI Montréal trip data.
Participants learn to import, clean, explore, and visualize data using the tidyverse.
Each step includes short explanatory text and auto-graded exercises, reinforcing understanding of the full data analysis cycle.

ggplot2 tutorial
This tutorial guides students through principles of the ggplot2 grammar of graphics.
It shows how to build effective and aesthetic visualizations, customize graphical elements, and interpret relationships between variables.
The interactive approach encourages direct experimentation and critical reflection on data visualization.

Overall objective:
provide reproducible and engaging learning environments where students learn through practice while receiving immediate feedback.
These tutorials are a key step in transitioning toward more interactive teaching, integrated with hybrid and online approaches.

Digital drawing with R project

Collaboration between art and data science exploring the graphical potential of R and the ggplot2 library.
Support: creation grant.
Objective: produce generative images based on statistical processes, document the creative functions, and share resources on GitHub.
Outcome: an interdisciplinary project at the intersection of art, science, and data visualization.

Outreach

I regularly share my practices in conferences and workshops focused on higher education pedagogy and data science.
I lead training sessions on:

  • scientific reproducibility with R and Quarto,
  • collaborative project management with GitHub,
  • design of interactive activities with Learnr.

These initiatives help build a francophone community of practice around data science and open digital pedagogy.

Here are examples of material on GitHub: