knitr::opts_chunk$set( collapse = TRUE, comment = "#>", fig.path = "man/figures/README-", out.width = "100%" ) library(areal) library(dplyr) library(sf) data(ar_stl_asthma, package = "areal") asthma <- ar_stl_asthma data(ar_stl_race, package = "areal") race <- ar_stl_race data(ar_stl_wards, package = "areal") wards <- ar_stl_wards
Areal interpolation is the process making estimates from a source set of polygons to an overlapping but incongruent set of target polygons. One challenge with areal interpolation is that, while the processes themselves are well documented in the academic literature, implementing them often involves "reinventing the wheel" by re-creating the process in the analyst's tool choice.
While the R
package sf
does offer a basic interface for areal weighted interpolation (st_interpolate_aw
), it lacks some features that we use in our work. The areal
package contains a suite tools for validation and estimation, providing a full-featured workflow that fits into both modern data management (e.g. tidyverse
) and spatial data (e.g. sf
) frameworks.
An article describing areal
's approach to areal weighted interpolation has been published in the The Journal of Open Source Software. The article includes benchmarking of areal
performance on several data sets. Please cite the paper if you use areal
in your work!
The easiest way to get areal
is to install it from CRAN:
install.packages("areal")
The development version of areal
can be accessed from GitHub with remotes
:
# install.packages("remotes") remotes::install_github("chris-prener/areal")
Note that installations that require sf
to be built from source will require additional software regardless of operating system. You should check the sf
package website for the latest details on installing dependencies for that package. Instructions vary significantly by operating system.
Two function prefixes are used in areal
to allow users to take advantage of RStudio's auto complete functionality:
ar_
- data and functions that are used for multiple interpolation methodsaw_
- functions that are used specifically for areal weighted interpolationThe package contains four overlapping data sets:
ar_stl_race
(2017 ACS demographic counts at the census tract level; n = 106) ar_stl_asthma
(2017 asthma rates at the census tract level; n = 106)ar_stl_wards
(the 2010 political subdivisions in St. Louis; n = 28). ar_stl_wardsClipped
(the 2010 political subdivisions in St. Louis clipped to the Mississippi River shoreline; n = 28). These can be used to illustrate the core functionality of the package. The following examples assume:
> library(areal) > > race <- ar_stl_race > asthma <- ar_stl_asthma > wards <- ar_stl_wards
areal
currently implements an approach to interpolation known as areal weighted interpolation. It is arguably the simplest and most common approach to areal interpolation, though it does have some drawbacks (see the areal weighted interpolation vignette for details). The basic usage of areal
is through the aw_interpolate()
function. This is a pipe-able function that allows for the simultaneous interpolation of multiple values.
In this first example, the total estimated population (TOTAL_E
) of each ward is calculated from its overlapping census tracts:
aw_interpolate(wards, tid = WARD, source = race, sid = "GEOID", weight = "sum", output = "sf", extensive = "TOTAL_E")
This example outputs a simple features (sf
) object and uses one of two options for calculating weights. All of these arguments are documented both within the package (use ?aw_interpolate
) and on the package's website.
What results from aw_interpolate()
is mapped below. Total population per census tract in St. Louis is mapped on the left in panel A. Using aw_interpolate()
as we did in the previous example, we estimate population counts for Wards in St. Louis from those census tract values. These estimated values are mapped on the right in panel B.
knitr::include_graphics("man/figures/exampleMap.png")
Both extensive and intensive data can be interpolated simultaneously by using both the extensive
and intensive
arguments. In this second example, the asthma and race data are combined, and estimates for both the population values and asthma rates are calculated for each ward from its overlapping census tracts:
# remove sf geometry st_geometry(race) <- NULL # create combined data race %>% select(GEOID, TOTAL_E, WHITE_E, BLACK_E) %>% left_join(asthma, ., by = "GEOID") -> combinedData # interpolate wards %>% select(-OBJECTID, -AREA) %>% aw_interpolate(tid = WARD, source = combinedData, sid = "GEOID", weight = "total", output = "tibble", extensive = c("TOTAL_E", "WHITE_E", "BLACK_E"), intensive = "ASTHMA")
Another advantage of areal
is that the interpolation process is not a "black box", but rather can be manually completed if necessary. Functions for validating data, previewing the areal weights, and walking step-by-step through the interpolation process are provided. See the areal weighted interpolation vignette for additional details about this workflow.
We are planning to experiment with at least three additional techniques for areal interpolation for possible inclusion into the package. These include:
We do not have a timeline for these experiments, though we are planning to begin experimenting with the pycnophylactic method in the coming months. We will be keeping the issues (linked to above) updated with progress. If you are interested in bringing these techniques to R
, please feel free to contribute to the development of areal
. The best place to start is bt checking in on our GitHub issues for each technique to see what help is needed!
Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.