retroharmonize
Ex Post Survey Data Harmonization
Survey harmonization tooling that spent its last release earning its way back onto CRAN.
◆Recent moves
- 3mo ago
CRAN modernisation release
Restores CRAN availability after the package was archived along with a dependency, on the back of substantial internal refactoring of the vctrs integration and a cleaner split between core, print, arithmetic, and vctrs methods. No new harmonization capability, but for a package whose users install from CRAN, being installable again is the change that matters.
View source ↗ - 7mo ago
Release candidate; notes cite code quality with no detail
A one-line entry claiming code quality improvements and specifying none. It reads as a staging step toward the CRAN restoration that followed four months later.
View source ↗ - 4y ago
Improved documentation
A three-line entry covering documentation, more consistent parameters, and reading STATA files. The parameter consistency is user-facing, though the notes are too thin to establish whether STATA support arrived here or was merely documented.
View source ↗ - 5y ago
0.1.7 With case study on harmonizing Arab Barometer surveys
The most substantive feature release in the window: harmonize_var_names() with suggestion helpers for systematic renaming across a list of surveys, subset_waves() for working with a single wave, perl-style regex in harmonize_values(), and a date-harmonization bug fix. The Arab Barometer vignette that ships alongside is the package arguing its case on real multi-country survey data.
View source ↗ - 5y ago
retroharmonize 0.1.13 is on CRAN
The first CRAN release, restating the package's goal rather than describing new work. It is a distribution milestone on top of the capability introduced two months earlier, and the moment the tooling became installable for the survey researchers it targets.
View source ↗ - 6y ago
0.1.0 Retrospective harmonization of survey data files in R
⚡ SPARKThe founding release, and the one that sets the approach every later version builds on: harmonization of value codes, value labels, and missing-value ranges carried through S3 classes so the transformations are recorded rather than applied and forgotten.
View source ↗