In MengLeZhang/socialFrontiers: What the Package Does (One Line, Title Case)

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(socialFrontiers)
library(dplyr)

Load the data from the package namespace using data()

The package contains an example dataset of London split into spatial polygons. Each polygon represents a census unit called lower super output area. The london data contains the number of UK-born and non-UK-born residents in a LSOA. For the purposes of this vignette, I select all areas in the borough of Barnet since London is so large.

The data is in the form of a sf object but the frontier detection function also accepts sp objects.

data(london)

##  Filter to the borough of Barnet
barnet <-
  london %>%
  filter(substr(LSOAname, 1, 6) %in% 'Barnet')

Using frontier_detect to find frontiers

Currently the only method for finding frontiers is the localised binomial model used in Dean et al. The model needs total counts of an event occuring (e.g. number of non-UK-born residents) and the total number of trials (e.g. total number of residents). The name of the variables denoting the column containing these counts must be entered as a string.

y <- 'nonUK' # 'nonUK' # Number of foreign
n.trials <- 'totalPop' #total population (per zone?)

Now we run the frontier_detect routine. You can see the underlying code used for the statistical model in using socialFrontiers:::binomial_localisedINLA.

frontier_model <-
  frontier_detect(
    data = barnet,
    y = y, n.trials = n.trials)

class(frontier_model) # Outputs a frontier_model object

The output saves as a 'frontier_model' object which can be used with other methods such as summary.

Methods for use with the frontier_model object

Summary

summary(frontier_model) ## This calls up summary.frontier_model

graphing and gis methods

We can extract the frontier (as well as non-frontier) borders as a sf object for further graphing or gis methods using frontier_as_sf. Note that this function throws up ignorable warnings that come from using sf::st_intersects.

suppressWarnings(borders_sf <-
                   frontier_as_sf(frontier_model, silent = T))

class(borders_sf)

Example graph using tmap showing all frontiers.

library(tmap)

##  Create a variable for prop-non-UK
barnet <-
  barnet %>%
  mutate(propNonUK = nonUK/totalPop)

tm_shape(barnet) +
  tm_fill(col= 'propNonUK') +
  qtm(borders_sf) # works