Skip to contents

PrjThaiHFID 0.2.0

Data-pipeline, documentation, and website release.

Anonymization

  • Hardened the data-raw/id_anonymize/ pipeline so no data-raw artifact carries true keys. anonymize_whitem160aggregate() now drops all true ID columns (hhid, hhid_month, province, village, household, villagehh), keeping only the anonymized id (tmid_hh).
  • Added anonymized-only crosswalks tm_key_id_hh_anony.rda and tm_key_id_vil_anony.rda; true-key crosswalks (tm_key_id*.rda) and *_true_id.csv are now git-ignored and stay local-only.
  • Repointed ffv_invest_return_bridge to the anonymized household crosswalk. Verified every vignette resolves only to anonymized data/ and data-raw/ files.

Output saving

  • Added exported helpers ffp_save_res_table() (LaTeX + optional CSV) and ffp_save_res_figure() (PNG + EPS) for writing vignette outputs to per-topic res/res_<topic>/ folders.
  • Each table/figure vignette now has a per-output ls_save_res control list (all default FALSE); generated res/ artifacts and inst/extdata/ exports are git-ignored and not shipped.

Vignettes and website

  • Migrated the eight analysis vignettes to Quarto (.qmd) and added Quarto setup to the pkgdown GitHub Action.
  • New pkgdown home page authored in README.qmd (rendered to README.md) with data-lineage, R-program-group, and vignette-tier Mermaid diagrams.
  • New Project Map navbar page (project_map.qmdproject_map.md) with program/data/data-raw/vignette/issue cross-reference tables and a run-order hierarchy diagram.
  • Added an explicit, dependency-ordered Articles section with per-vignette descriptions.
  • Reconciled the reference index to the committed pipeline: removed three non-vignette datasets (tstm_hh_assets_mth25, tstm_hh_attributes_mth25, tstm_i2l2b) and registered the new ffp_save_res_* helpers.

Documentation

  • Added GitHub-issue and vignette cross-references throughout the data (R/data.R, R/data-res.R) and function roxygen, and created a tracking issue for ffv_loan_overlap.