Create a custom proximity score for some user-specified set of points, so that every US block group is assigned a proximity score counting those points

Indicator of proximity of residents in each US blockgroup to a custom set of facilities or sites

Usage

proxistat(
  topoints,
  bpoints = blockpoints[state_from_blockid(blockpoints) == "DE", ],
  blocks_per_batch = 1000,
  countradius = 3.106856,
  maxradius = 621.3712,
  quadtree = NULL,
  quaddatatable = NULL
)

Arguments

topoints: data.table of lat lon
countradius: distance within in which nearby sites are counted to create proximity score. In miles, and default is 5 km (5000 / meters_per_mile = 3.106856 miles) which is the EJScreen zone for proximity scores based on counts.
maxradius: max distance in miles to search for nearest single facility, if none found within countradius. EJScreen seems to use 1,000 km as the max to search, since the lowest scores for proximity scores of RMP, TSDF, or NPL are ROUGHLY 0.001, (exactly 0.000747782) meaning approx. 1/1000 km and km_per_mile = 1.609344 = meters_per_mile / 1000 so 1000 km is 1000 / 1.609344 = 621.3712 miles. However, the exact min value implies 1337.288 kilometers, or 830.9523 miles?
quadtree: Index of sites such as facilities that will be the basis for the proximity scores. Optional, because it can be created here on the fly based on pts parameter, but can pass it if already exists - an index of block locations, built during use of EJAM package. create_quaddata()
quaddatatable: optional, created from pts if not passed, created by create_quaddata() utility, and used to create quadtree

Value

data.table of block groups, with proximityscore, bgfips, lat, lon, etc.

Details

A "facility" proximity score for the residents in one place is an indicator of how far away those facilities are, and how many there are nearby - it accounts for the number of facilities within 5 kilometers (facility density) and the distance of each (proximity). If there are more points nearby, and/or the points are closer to the average resident in a blockgroup, that blockgroup gets a higher proximity score.

The formula for this proximity score is the sum of (1/d) where each d is distance of a given site in kilometers, summed over all sites that are within 5 km (or the single closest site if none are within 5 km), just as in EJScreen proximity scores like the TSDF or RMP scores.

Any custom user-provided set of points can be turned into a proximity score, such as locations of all industrial sites of a certain type, or all grocery stores, or all schools. A proximity score can be created for all blocks and block groups in the US (or just one State or Region). Then the proximity scores can be analyzed in a tool like EJAM, just as the existing pre-calculated proximity scores are analyzed to represent the number of nearby hazardous waste treatment stoprage and disposal facilities, weighted by how far away each one is, as provided in the EJScreen proximity score for TSDFs.

A custom user-specified proximity score might focus on schools, for example. The schools proximity score could be analyzed in EJAM for one or more communities, or areas near regulated facilities, or any set of analyzed places. That would provide statistics demonstrating which places have more schools closer to them (or inside the areas defined by polygons or FIPS codes, for example).

To create the proximity score, EJAM uses the same method EJScreen used to create proximity scores. The specified points first get indexed by a utility function called indexpoints() and are searched for and counted near every block and blockgroup in the US via a function called getpointsnearby().

Examples

 # pts <- testpoints_100
 # x <- proxistat(topoints = pts[1:1000,])
 #
 # summary(x$proximityscore)
 # # analyze.stuff   pctiles(x$proximityscore)
 #
 # plot(x = x$lon, y = x$lat)
 # tops = x$proximityscore > 500 & !is.infinite(x$proximityscore) & !is.na(x$proximityscore)
 # points(x = x$lon[tops], y = x$lat[tops], col="red")