replicateEverything: Reproduce Empirical Research Results

The replicateEverything package provides tools for discovering and executing computational replications of empirical research papers. It connects to a public replication registry containing metadata, replication scripts, and processed datasets required to reproduce figures and tables from published studies.

Workflow

A typical workflow using the package is:

1. Browse the registry with load_index() or search_papers().

2. Inspect available replications using list_replications().

3. Run a single table or figure with run_replication(doi, "fig_1").

4. Reproduce all results with run_replication(doi, "everything").

5. View replication code with get_code().

6. Launch the bundled Shiny demo with run_shiny_app(), or deploy it with save_local_shiny().

7. Contribute a study with build_study_outputs(), check_and_bake_study(), and register_study() (maintainer).

8. Refresh derived registry files with refresh_registry(), inspect health with audit_report(), or run a live audit with audit_everything().

See vignette("meet-the-functions") for a tour of every main function.

Shiny demo

A live instance runs at https://shiny2.wzb.eu/ipi/replicate/. The package ships a demo app in inst/shiny/. Use run_shiny_app() to launch it from an installed build, or save_local_shiny() to copy app.R and www/ into a Shiny Server directory.

Registry

Replication metadata and materials are stored in the public registry: https://github.com/replicate-anything/registry.

See also

Useful links:

• https://github.com/replicate-anything/replicateEverything

• https://replicate-anything.github.io/replicateEverything/

• Report bugs at https://github.com/replicate-anything/replicateEverything/issues

Author

Maintainer: Macartan Humphreys macartan.humphreys@wzb.eu

Authors:

• Macartan Humphreys macartan.humphreys@wzb.eu

• Vermon Washington vermon.washington@wzb.eu

• Cord Masche cord.masche@wzb.eu