Credit Risk with R

This repository contains the source files for Credit Risk with R, a Quarto book by Dr. Martin Lozano. The book develops credit-risk models with reproducible R code, numerical examples, tables, and graphics.

Published site: https://mlozanoqf.github.io/tutorial_arf/

Scope

The current version focuses on:

• Logistic credit scoring, predicted PDs, cutoff rules, bad rates, ROC/AUC, calibration, and lending decisions
• Tree-based credit scoring with a single decision tree and XGBoost
• Structural credit risk with the Merton model, risk-neutral valuation, and model-implied default probabilities
• Credit spreads, corporate bond valuation, and CDS valuation in discrete time
• Portfolio credit risk with Gaussian copulas, t-copulas, correlated defaults, portfolio losses, Credit VaR, expected shortfall, and economic capital

Book Structure

• index.qmd: preface, publication metadata, and edition notes
• 01-logistic.qmd: logistic credit scoring
• 02-tree-based.qmd: tree-based credit scoring
• 02-merton.qmd: the Merton model
• 04-credit-spreads-cds.qmd: credit spreads and CDS valuation
• 03-gaussian-copula.qmd: copula models for portfolio credit risk
• references.qmd: references

The book configuration lives in _quarto.yml.

Repository Layout

• R/: helper scripts for data preparation, model estimation, formatting, and chapter setup
• loan_data_*.rds: loan data used in the credit-scoring chapters
• _book/: generated HTML output created by quarto render
• .github/workflows/publish.yml: GitHub Actions workflow for rendering and deploying the book to GitHub Pages
• styles.css and *.html partials: custom layout, navigation, progress, and page behavior
• _extensions/quarto-ext/fontawesome/: local Quarto extension used for icons
• references.bib: bibliography

Legacy .Rmd files and media assets are kept in the repository, but the active book build is driven by the .qmd files listed in _quarto.yml.

Render Locally

Install Quarto and R, then render the book from the repository root:

quarto render

For interactive local preview:

quarto preview

The GitHub Actions workflow installs the R packages needed for deployment. For local rendering, the main package set is:

• dplyr
• gmodels
• ggplot2
• knitr
• mnormt
• plotly
• pROC
• rmarkdown
• rayshader
• scatterplot3d
• Sim.DiffProc
• tidyr
• vembedr
• viridis
• xgboost

Publication

Pushing to main triggers .github/workflows/publish.yml. The workflow:

1. Checks out the repository.
2. Installs Quarto.
3. Installs R 4.5.2.
4. Installs the required R packages.
5. Runs quarto render.
6. Uploads _book/ as a GitHub Pages artifact.
7. Deploys the artifact to GitHub Pages.

Maintenance Notes

• Edit the source .qmd files and helper scripts, not generated files in _book/.
• Keep _quarto.yml synchronized with chapter additions, removals, or renames.
• If custom navigation changes, review styles.css and the HTML partials included in _quarto.yml.
• If a chapter needs new R packages, update .github/workflows/publish.yml so GitHub Pages can render the book.
• Do not commit large local reference PDFs unless they are intentionally part of the public repository.

License

This project is licensed under the GNU General Public License v3.0. See LICENSE.