loud-study

Research & Analytics used for the LOUD Study (MPI: Joudrey & Kolak) Student Research Assistant Support: Sam Wang (2026).

Overview

This repo stores data and analyses to generate a multivariate access index to localize opioid use response in the U.S., incorporating both conceptually-driven and data-driven approaches.

Organizations

• indicators_raw: Stores raw and/or mid-process variables that will be integrated in the final analyses.
• data_final: near-final versions of the data, including the important DataDictionary.csv file. The full dataset was too large to be stored on Github, but can be accessed by team members here.
• data-driven: includes working result outputs of PCA and K-means, and early versions of summary statistics.
• scripts: Includes R scripts for calculation of each stage (for conceptual indices), crosswalk to zip, and other wrangling or summary stats scripts.

Team members, please be sure to update the DataDictionary.csv and loud-cleaned.geojson files in both Git and the Google Drive Data Analysis folder to ensure we're aligned.

Project management is being split across Google Drive and Notion; see Marynia for questions.