stacked
  • Home
  • Method
    • Stacking, aggregation issues, and corrective weights
    • Compositional balance
    • Worked example: the sign flip
  • Interactive app
  • Vignettes
    • Introduction
    • Kappa trade-offs
    • Repeated cross-sections
    • Propensity-score weights
    • Applied example
    • Larger-than-memory (DuckDB)
  • Data
    • All datasets
    • medicaid
    • smoke
    • acs_micro
    • signflip
    • unparallel
  • Paper
  • Functions

Interactive app

This interactive app has a few defaults to show the corrective weights in action. You can also change the parameters yourself if you scroll down. The app runs entirely in your browser via webR. The first load can take a minute or two; so give it a moment to boot before interacting.

Open in its own tab ↗

Run it locally in R

If you have the package installed, the same app launches locally with:

library(stacked)
runInteractiveExample()

Regenerating the static export

The in-browser version is produced by vendoring build_stack() into the app and exporting it with shinylive. From the R package root:

Rscript dev/export-shinylive.R /path/to/stacked_did_website/app
 
Wing, Hollingsworth & Freedman — stacked difference-in-differences