cataScience 2.2.0

Added a shared two-cat navbar icon and favicon based on the maintainer’s cats, with a larger circular mark for both app launchers. The website uses the same cat favicon and retains its original hex logo.
Added run_catas() for an extended workshop; run_cata() remains the original entry point. The workflow connects review, audited joins, explicit denominators, charts and communication, with input snapshots and undo.
Added reusable import, profile, missingness, key-check, outlier-flag, join, proportion and status-summary functions, plus documented examples.
Added cata_plot() with 11 chart types and theme_cata(), and offline HTML report/dashboard exports. Project exports include reproducible R, Quarto and Shiny sources, input snapshots and preparation records.
Added Why data matters, John Snow’s map, a care-cascade lesson and the synthetic Lao workshop B/C analysis. Kept AI chart comparison as an external classroom exercise.
Kept the mice cat Easter egg and added multiple imputation, diagnostic chains and pooled model estimates using synthetic cats.
Removed rio. Made VIM optional for kNN; replaced its missingness chart with mice. Kept naniar/skimr and the optional DSIR theme.
Fixed classic outlier removal so source columns beginning with shadow_ are preserved. Added explicit UTF-8 loading for nested app scripts.
Revised quiz explanations to distinguish review flags from errors and missingness assumptions from conclusions established by observed data.

cataScience 2.1.4

Added three vignettes covering installation and launch, a reproducible data-quality workflow, and guidance for running a training session.
Added a pkgdown documentation website, linked from the package metadata and README, with a function reference and searchable articles.
Added a package hex logo based on the maintainer’s two cats, displayed in the README, website and vignettes.

The markdown lesson pages are now rendered from text instead of from the file path. markdown::mark() (>= 2.0) rewrites an input file in place when it runs under R CMD check, which made the test suite fail on CRAN’s Debian machines where the installed package library is mounted read-only (and violated the CRAN policy against writing outside the session temp directory). Rendering from text writes nothing; the generated HTML is unchanged.

Added a testthat test suite covering the app code that R CMD check cannot reach through the examples: the data-cleaning helpers (outlier detection, winsorizing, imputation, file import), the quiz and cleaning module servers via shiny::testServer(), the question bank, and a smoke test that builds the whole app UI.
The example of run_cata() now also runs outside an interactive session.
The outlier boxplot no longer warns about missing values while a dataset still has them.

Images inside the markdown lesson pages now scale to the card width instead of being clipped (affected the Background tabs of the Visualize, Cleaning, and Import pages).
The app stylesheet link carries a cache-busting version query so browsers pick up styling changes after an update.
Resized one oversized bundled figure (9000 px wide) to 1800 px.

First release: run_cata() launches the bundled “A Journey of Data Science” training app.