Skip to content

Latest commit

 

History

History
58 lines (48 loc) · 2.9 KB

README.md

File metadata and controls

58 lines (48 loc) · 2.9 KB
Raw

R build status Project Status: Active

US Hardship Index

us-hardship-index is an R package for calculating a hardship index for specified US states, using data from the US Census Bureau’s American Community Survey. It includes the single primary function hs_hardship_index(), which accepts the two parameters of:

  • “state” for the desired US state; and
  • “year” for the desired year (with values available since 2010).

The index itself is calculated from the methodology of “An Update on Urban Hardship,” by Lisa M. Monteil, Richard P. Nathan, and David J. Wright (2004), The Nelson A. Rockefeller Institute of Goverment. It is formed as a multiple of the following six measures, each standardised to unit (or percentage) scales:

  • occupancy: Proportion of rooms with > 1 occupant per room;
  • poverty: Proportion of households below poverty line;
  • unemployment: Proportion of unemployed adults;
  • no_hs: Proportion of population without highschool diploma
  • deps: Proportion of population who may be considered dependent; that is either under 18 or over 65
  • income: Per-capita income

All variables are quantified such that lower values are better, except for income. The hardship index is then simply the product of those six metrics, again standardised to a unit (or percentage) scale. An example of the index in action is provided by the city of Chicago, as a table of the six metrics plus their conversion to a composite hardship index.

Note that those Chicago data are ultimately transformed into a single rank for each measured area, and so manifest a perfectly uniform distribution. In contrast, the values returned by the hs_hardship_index() function are not transformed, and so generally manifest highly skewed distributions. The logarithm of these values will nevertheless generally be approximately normally distributed. Accordingly, any statistical analyses of hardship values should generally be applied to log-transformed versions of the values derived here.

Access to Census Bureau Data

This package requires an API key for census.gov, which can be obtained from the Census Bureau’s website. This key should be stored as an environment variable named CENSUS_API_KEY, generally by specifying its value in the ~/.Renviron file. Alternatively, the key can be set with the tidycensus function, census_api_key().