Replicate Everything

Output
Code

Contribute a replication

Hover underlined terms for examples.

Add a folder under papers/

Folder layout

                      papers/10.1017S0003055403000534/
  replication.yml
  code/tab_1.R
  data/repdata.dta
  artifacts/tab_1.rds
  artifacts/manifest.json

in the registry repository.

Include replication.yml

Sample replication.yml

                      paper:
  doi: https://doi.org/10.1017/S0003055403000534
  title: Ethnicity, Insurgency, and Civil War
  dependencies:
    - haven
    - modelsummary

replications:
  - id: tab_1
    type: table
    label: Table 1
    data: data/repdata.dta
    code: code/tab_1.R
    format: format_tab_1
    artifact: artifacts/tab_1.rds

, code/

Self-contained script

                      # Paper folder: https://github.com/replicate-anything/registry/
#   tree/main/papers/10.1017S0003055403000534

library(haven)
library(modelsummary)
library(kableExtra)

make_tab_1 <- function(data) { ... }
format_tab_1 <- function(object) { ... }

make_tab_1(haven::read_dta("../data/repdata.dta")) |> format_tab_1()

, and data/ .

Prefer two steps when possible: make_tab_1(data)

Analysis function

                      make_tab_1 <- function(data) {
  glm(onset ~ warl + gdpenl + lpopl, data = data,
      family = binomial())
}

format_tab_1 <- function(object) {
  modelsummary::modelsummary(
    object,
    output = "kableExtra",
    stars = TRUE
  )
}

for analysis and format_tab_1(object)

Display function

                      make_tab_1 <- function(data) {
  glm(onset ~ warl + gdpenl + lpopl, data = data,
      family = binomial())
}

format_tab_1 <- function(object) {
  modelsummary::modelsummary(
    object,
    output = "kableExtra",
    stars = TRUE
  )
}

for display. Register with format: format_tab_1

replication.yml entry

                      paper:
  doi: https://doi.org/10.1017/S0003055403000534
  title: Ethnicity, Insurgency, and Civil War
  dependencies:
    - haven
    - modelsummary

replications:
  - id: tab_1
    type: table
    label: Table 1
    data: data/repdata.dta
    code: code/tab_1.R
    format: format_tab_1
    artifact: artifacts/tab_1.rds

so artifacts store analysis output (usually .rds ). Omit format when analysis and formatting cannot be separated. Figures use make_fig_1(data)

Figure example

                        - id: fig_1
    type: figure
    data: data/example.csv
    code: code/fig_1.R
    artifact: artifacts/fig_1.png

make_fig_1 <- function(data) {
  ggplot2::ggplot(data, ggplot2::aes(x, y)) +
    ggplot2::geom_line()
}

and save .png artifacts.

Scripts must be self-contained

Script pattern

                      # Paper folder: https://github.com/replicate-anything/registry/
#   tree/main/papers/10.1017S0003055403000534

library(haven)
library(modelsummary)
library(kableExtra)

make_tab_1 <- function(data) { ... }
format_tab_1 <- function(object) { ... }

make_tab_1(haven::read_dta("../data/repdata.dta")) |> format_tab_1()

: include library() calls, a commented link to the paper folder on GitHub, and a footer that loads data/ via a local path (works when run from code/ or the paper folder).

Run Rscript scripts/build_artifacts.R

Build & validate

                      cd registry
Rscript scripts/build_artifacts.R

# one paper only:
Rscript scripts/build_artifacts.R papers/10.1017S0003055403000534

Rscript scripts/validate_artifacts.R

in the registry to precompute artifacts/ for the app.

1. Choose a study

2. Tables & figures

Contribute a replication