This repository provides a set of scripts to help make use of the Social Connectedness Index (SCI) data. The SCI data are downloadable at: https://data.humdata.org/dataset/social-connectedness-index.

It also includes replication code for Kuchler, Russel, and Stroebel 2021.

You can find the replication code for another paper that uses the SCI (Bailey, Kuchler, Johnston, Russel, State, and Stroebel 2020) in this separate repository.

Separately, we host the zip file gadm_based_shapefiles.zip which contains a set of shapefiles (in .shp format and, for R users sf objects in .Rds files) that can be matched to the SCI data for mapping. These are built from the shapefiles for GADM version 2.8 and European NUTS 2016 (see sources and their relevant terms of use below). The zip also includes html files with interactive maps that can be used to explore the shapes.

IMPORTANT NOTE: This repository uses git-lfs for versioning large files. You will need it installed to clone the repository.

The resources are split into 2 main directories:

1. example_scripts contains a set of example scripts in R that map the SCI data. It includes subfolders for each of the different SCI granularities (country_country, county_county, etc.). It also includes an interactive_map subfolder that provides example code to generate interactive html maps using the Leaflet R package. You can view the example maps by downloading the html file and opening it in any internet browser (e.g. Google Chrome). An example interactive SCI map is hosted here.

2. covid19_exploration contains a set of example scripts in R and Stata that produce the results in Kuchler, Russel, and Stroebel 2020. This folder contains a separate short readme.

We also include Relevant Literature + Bibtex.bib, a list of papers that introduce and develop the Social Connectedness Index, as well as guidance on how to cite the prior literature when using the SCI data.

To generate the results in covid19_exploration, we use a number of data from a number of sources:

1. ACS_17_5YR_DPO5.csv are county-level demographics from the American Community Survey.

2. cty_covariates_oi.csv are additional county-level demographics from Opportunity Insights.

3. NCHSURCodes2013.csv are National Center for Health Statistics Urban-Rural County Classifications.

4. sf12010countydistancemiles.csv are county-to-county distances from the National Bureau of Economic Research.

5. The eurostat folder contains European NUTS3 region demographics, made available by Eurostat.

6. COVID data are pulled directly from Github repositories hosted by Johns Hopkins University and Dipartimento della Protezione Civile.

To generate the shapefiles in gadm_based_shapefiles, we bring together two sets of shapefiles:

1. For European NUTS2 and NUTS3 regions, we use (c) EuroGeographics for the administrative boundaries, available here.

2. For non-European countries, we use the version 2.8 Database of Global Administrative Areas (GADM) shapefiles, available here.

This repository is managed by Theresa Kuchler and Johannes Stroebel.