cliarappak

The goal of cliarappak is to package the Country-Level Institutional Assessment and Review (CLIAR) Benchmarking Dashboard as an installable, reproducible golem Shiny application. It replaces the original script-style cliarapp project with a proper R package, while keeping the dashboard itself unchanged for end users. It provides the following features:

• Country Benchmarking: Compares institutional indicator performance for a base country against comparator countries or predefined groups using closeness-to-frontier (CTF) scores.

• Modular Architecture: Each dashboard tab is a self-contained Shiny module (mod_*.R), wired together through a single golem app_ui() / app_server() pair. Shared reactive state is passed between modules as explicit arguments.

• Reproducible Data: All indicator data is sourced at runtime from the cliaretl package — no data files are bundled in this repository.

• One-Command Deployment: deploy_app() regenerates the Posit Connect entry file with the current runtime options baked in, then publishes it, so a chosen data vintage survives Connect’s own restart and scale-out schedule.

Project Structure

The package follows the standard golem layout:

cliarappak/
├── R/                        # Package source
│   ├── run_app.R             # Launches the Shiny application
│   ├── app_ui.R              # Top-level golem UI (bs4Dash dashboard shell)
│   ├── app_server.R          # Top-level golem server (wires the modules)
│   ├── app_config.R          # app_sys() and golem config helpers
│   ├── fct_app_data.R        # Builds the shared data objects from cliaretl
│   ├── fct_*.R               # Pure helper / data-preparation functions
│   ├── mod_*.R               # One Shiny module per dashboard tab
│   ├── guides.R              # Cicerone guided-tour step definitions
│   └── deploy_app.R          # Posit Connect deployment wrapper
├── inst/
│   ├── app/www/              # Static assets (CSS, images, PPTX/DOCX templates)
│   ├── rmd/                  # Word and coverage report templates
│   ├── extdata/              # publicationsList.xlsx
│   └── golem-config.yml      # golem configuration
├── dev/                      # golem development scripts (run_dev.R)
├── man/                      # Auto-generated R documentation (via roxygen2)
├── tests/                    # Unit tests (testthat)
├── renv/                     # Dependency management (via renv)
├── DESCRIPTION               # Package metadata and dependencies
├── NAMESPACE                 # Package namespace (exports / imports)
├── renv.lock                 # Lockfile for reproducible package versions
├── README.Rmd               # Source file for README.md
└── README.md                # Project documentation (rendered version)

Dashboard Tabs

Tab	Description
Country Benchmarking	Closeness-to-frontier scores for a base country vs. comparator countries or groups, with static and dynamic (time-varying) views
Cross-Country Comparison	Indicator-level bar charts across the selected countries
Bivariate Correlation	Scatter plots of two indicators with an optional linear fit
World Map	Choropleth map of CTF scores or original indicator values
Time Trends	Year-on-year changes in raw indicator values
Data	Browsable data table with Excel / CSV / Stata download
Methodology & User Guide	Methodology documentation, user guide, and downloads
Publications	Country-filterable list of CLIAR-related publications

Package Installation

Install the cliarappak package with:

# Step 1. Install the packages 'pak' or 'remotes' if you don't have them:

# install.packages("pak")
# install.packages("remotes")


# Step 2. Install cliarappak from GitHub:
remotes::install_github("WB-PIDA-Data-Science-Shop/cliarappak")

This also installs cliaretl from GitHub (declared in Remotes:), which provides all of the dashboard’s data.

Running the App

library(cliarappak)

# Launch the dashboard
run_app()

# Preview a specific dynamic-benchmarking data vintage
run_app(dynamic_year_cutoff = 2022)

For development, restore the locked environment and use the golem workflow:

renv::restore()
devtools::load_all()
golem::run_dev()

Deployment

The dashboard is deployed to Posit Connect via deploy_app(). It reads the target application GUID from an environment variable (cliarappak_dev_guid or cliarappak_prod_guid, set in .Renviron or via Sys.setenv()):

# Development slot
deploy_app(type = "dev")

# Production slot, with the auto-derived current-cycle cutoff
deploy_app(type = "prod")

# Ship a specific data vintage
deploy_app(type = "prod", dynamic_year_cutoff = 2022)

Data

All data is provided at runtime by the cliaretl package, including:

• closeness_to_frontier_static / closeness_to_frontier_dynamic — CTF scores
• db_variables_final — indicator metadata
• wb_country_groups / wb_country_list — country and group reference tables

Reproducibility

The cliarappak package uses renv to lock package versions, ensuring consistent results across environments. To set up the project environment for development, run renv::restore().