Advanced Statistical Modelling (STATS 720) — Portfolio

Interactive analyses and write-ups for STATS 720 (Statistical Modelling), authored in Quarto and rendered as a website.

• Live site: https://KostasBanos.github.io/Advanced_Linear_Models/
• Example page: https://KostasBanos.github.io/Advanced_Linear_Models/Project_4.html

The website provides interactive HTML outputs (e.g., Plotly figures with zoom/hover) so you can explore results without running code locally. PDF renders are also included for archival and offline reading.

---

Overview

This repository contains Quarto notebooks (.qmd) developed for STATS 720 at McMaster University (instructor: Prof. Ben Bolker). The work emphasizes:

• rigorous statistical reasoning and communication
• modern modelling workflows in R with reproducibility
• principled model comparison and diagnostics

Topics include GLMs, LMMs/MLMs, GLMMs, bias-reduced/penalized/Bayesian estimation, simulation-based inference and bootstrap methods, residual diagnostics, spline-based nonlinear effects, and likelihood/resampling-based comparisons.

---

Requirements

• R (≥ 4.1) and RStudio (current recommended)
• Quarto (install from https://quarto.org or via RStudio)

You may need to install the R packages used in each project. Open a .qmd file to see the libraries loaded at the top and install any that are missing.

Methodological Scope

Across the projects, the following modeling frameworks and techniques are explored:

• Generalized Linear Models (GLM)
• Linear and Multilevel Mixed Models (LMM / MLM)
• Generalized Linear Mixed Models (GLMM)
• Bias-reduced, penalized, and Bayesian estimation
• Simulation-based inference and bootstrap methods
• Model diagnostics and residual analysis
• Nonlinear effects and spline-based modeling
• Likelihood-based and resampling-based model comparison

---

Quick start

Option A: RStudio

1. Open the RStudio project: stats720-portfolio.Rproj.
2. Render a single notebook (open a .qmd and click “Render”) or run:

quarto::quarto_render("Project_1.qmd")

3. Render the entire site to docs/ (for GitHub Pages):

quarto::quarto_render()

Option B: Terminal (Quarto CLI)

From the project root:

quarto render            # build all pages into docs/
quarto render Project_1.qmd
quarto preview           # local dev server with live reload

---

Project structure

• index.qmd — site landing page
• Project_*.qmd — notebooks for Projects 1–5
• docs/ — site output (committed for GitHub Pages)
• _quarto.yml — site config (navbar, theme, output dir)
• data/olymp1.csv — input data used in Project 1
• Project_Files/ and docs/*_files/ — figures/assets generated during rendering
• stats720-portfolio.Rproj — RStudio project file
• LICENSE, CITATION.cff, README.md

---

Publishing (GitHub Pages)

This site is configured to render into docs/. To publish:

1. Build the site (RStudio “Render” or quarto render).
2. In your GitHub repository settings, set Pages → “Deploy from a branch” and select the main (or default) branch with /docs as the folder.
3. Visit the GitHub Pages URL after it deploys.

---

Citation

If you use or reference this work, please cite it. See CITATION.cff for full metadata.

Example (APA-like):

Banos, K. (2026). STATS 720 — Data Science Portfolio (v1.0.0) [Software]. MIT License. https://github.com/KostasBanos/Advanced_Linear_Models

---

Acknowledgements

Course: STATS 720 — Statistical Modelling, McMaster University.
Instructor: Prof. Ben Bolker.

---

License

MIT — see LICENSE.