nearpresence

The nearpresence package implements Near Presence Cluster Analysis, a statistical method for identifying spatial clustering within presence / absence data observed in irregularly distributed areal observation units. It was developed to aid in the interpretation of data from archaeological surface survey, specifically, to help identify clusters of units containing material from a certain chronological period. However, it will be useful in any context where binary observations are recorded in areal units that are irregularly distributed and discontiguous.

Example

This example comes from the Molyvoti, Thrace, Archaeological Project.
The following plot shows the distribution of units with Classical-period material in a portion of the survey area:

library(nearpresence)
library(sf)
#> Warning: package 'sf' was built under R version 4.0.5
data(tracts)
data(chron)
tr.chr<-merge(tracts, chron, by.x = "UnitID", by.y = "Survey_Uni")
par(xpd=TRUE)
plot(tr.chr["Clas"])

Spatial structure is difficult to discern. Near Presence Cluster Analysis identifies units that have Classical-period material present and have more neighbors with Classical-material than would be expected under conditions of complete spatial randomness.

data(tracts)
data(chron)
np<-NP(chron = chron,
   chron_ID = "Survey_Uni",
   periods = "Clas",
   tracts = tracts,
   tracts_ID = "UnitID",
   swl = IDW_nnear(tracts = tracts, tracts_ID = "UnitID", n = 8),
   perms = 100,
   cut = 0.05)
#> ================================================================================
par(xpd=TRUE)
plot(np["Clas_Res"], col=NP_colors(np[["Clas_Res"]]))
legend("topleft", 
       legend = c("Present, High NP", "Present, Low NP", "Present, Moderate NP", "Absent, High NP", "Absent, Low NP", "Absent, Moderate NP"), 
       fill = c("#e60000", "#feff73", "#ffaa01","#febebe", "#ddd2ff", "#b5d69f"),
       inset = c(-0.1, -0.1))

Overview

The nearpresence package contains four functions, which are meant to be used consecutively:

1. IDW_nnear and IDW_radius create a spatial weights model in the form of a list from a sf object representing the spatial distribution of observation units.
2. NP uses the output of either IDW_nnear or IDW_radius and a table with chronological data to perform NPCA. It returns a sf object with various results of NPCA stored as attributes. 3. NP_colors is useful for displaying the output of NP by applying a pre-determined color scheme.

Installation

You can install the development version of nearpresence from github with:

install.packages("devtools")
devtools::install_github("eweaverdyck/nearpresence")