Welcome to our repository for the physician-to-voter linkage and subsequent analyses.
This repository is organized as follows:
.
├── code
│ ├── 00_unzip_l2.R
│ ├── 01_extract_l2.R
│ ├── 03_clean_physician_data.R
│ ├── 04_locality_sensitive_hash.R
│ ├── 05_match_model.R
│ └── match_diagnostics.R
├── figures
│ ├── age_dist_hist.png
│ ├── at_a_glance.png
│ ├── matches_by_state.png
│ ├── processing.png
│ └── screenshot.png
├── linkage_slides.Rmd
├── pres.bib
├── README.md
└── _targets.R
-
The
code/
directory houses the functions used to complete the linkage and subsequent diagnostics. -
The
_targets.R
file contains the master script to run the entire analysis pipeline using the targets R package, which rebuilds downstream pieces of code if and only if uspstream code has changed. This saves time and means we can ensure reproducibility without re-running the pipeline after each change. The pipeline can be run by callingtargets::tar_make()
in R. -
The
figures/
directory holds the outputs produced by the pipeline. A seperatetables/
folder will likely be created further along in the project. -
linkage_slides.Rmd
andpres.bib
are both used to generate the slides shared with the team.
Here is a path of the code dependencies in the repository. It is not live-updating, and so will not respond to changes in the architecture / pipeline, but it is useful to see a roughly-current summary of the analysis code now.
Package site for zoomerjoin, our in-house package to perform fast fuzzy-linking on name.