NEWS
sta310r 0.1.0
- The package is now published on R-universe at
https://eribul.r-universe.dev, which serves pre-built binaries for Windows,
macOS and Linux so installation no longer requires a local toolchain (for
example Rtools on Windows).
- Added a new pair of vignettes, "Computer exercise 1: Describing and
comparing survival data" and its solutions, restricted to Chapters 1–2 of
Collett (2023). The exercise covers censoring, the survivor/hazard/cumulative
hazard functions, the empirical, Kaplan–Meier and Nelson–Aalen estimators,
Greenwood intervals, hazard estimation, and the log-rank, Wilcoxon,
Peto–Peto, stratified and trend tests. The life table is not covered. The
exercise uses only data sets included in the package and does not cover Cox
regression. The exercise and solutions are generated from a single source
file; the
show_solutions parameter controls whether the solution is
included immediately after each question.
- Removed the teaching functions and their documentation and tests. The
package now focuses on the data sets and the computer exercises.
- Added
dataset_by_exercise, an index that links every data set to the worked
examples of Collett (2023) that use it, with the example number, chapter,
example title and the data set dimensions.
- Every data set that is used in a numbered example now has an alias
dX.Y,
where X.Y is the example number. The aliases are ordinary data objects, so
d3.7 is the myeloma data used in Example 3.7 and d8.4 is the baseline
cirrhosis data. A data set used in several examples has one alias per
example. The aliases are listed on the help page of the corresponding data
set.
- Added the vignette "Data sets by exercise", which presents the index, and an
overview of it in the README.
- The examples on the help page of every data set now reproduce the analyses
that Collett (2023) presents with that data set, with a comment for each
example number. Each page has two parts: first the code of Chapter 17 (the
survival package, penalized, eha, cmprsk, and so on), then the same
analysis with more recent tools (ggsurvfit, flexsurv, coxme, icenReg,
tidycmprsk, broom, rms, glmnet). Data sets that are not used in a
numbered example show the analyses suggested for the appendix or, for
scottish_mbc_cohort, the modern tools only.
- First release: the data sets from Collett (2023) and the computer exercises
used in the STA310 course.
- Added the
scottish_mbc_cohort data set, the synthetic Scottish metastatic
breast cancer cohort that is used in the first computer exercise. It does not
come from Collett (2023); its help page gives the source and licence.
- Added a "Getting started with sta310r" vignette, a "Computer exercise 1:
Survival analysis" vignette and a "Computer exercise 1: solutions" vignette.
The exercise and its solutions are separate documents that share the same
question numbering.
- The vignettes are shipped as pre-built HTML in
inst/doc/, so
vignette(package = "sta310r") works also when the package is installed
without building vignettes (dev/refresh-vignettes.sh updates them).
- Added unit tests for the structure of the data sets.