wrobertlong / nearrepeat Goto Github PK

The R package NearRepeat uses the Knox test for space-time clustering to quantify the spatio-temporal association between events. In criminology, this test has been used to identify people and locations that are at disproportionate risk of victimization. Of interest is often not only the ‘same repeat’ victims (people or locations that after a first crime event, are targeted again within a short period thereafter), but also the ‘near repeat’ victims: nearby people or locations that are victimized within a short period after the first crime. For more information, see for example the JDI Analysis Brief.

Note that in parallel with the development of the code here, Toby Davies has developed an alternative implementation in Python, which can be found here. We have collaborated on some of the accompanying materials and reviewed each other’s code, and as far as we are aware the two implementations should give identical results.

Several other options exist but have drawbacks:

• Prof. Jerry Ratcliffe’s / Temple University’s Near Repeat calculator only works on Windows (and needs admin rights to execute and write files). It seems this software mislabels the rows and columns of its output files, which may lead to erroneous conclusions.

• the free JDI Near Repeat Toolkit also only works on Windows. However it is not actively supported and does not run on Windows 10.

• the R package surveillance has implemented the knox() function, but this only works for a single spatial distance and temporal distance, whereas the current package also works with multiple bands of spatial and temporal distances, as well as same repeat analysis.

The most recent version of the package is:

Steenbeek, W. (2018). Near Repeat. R package version 0.1.0. URL: https://github.com/wsteenbeek/NearRepeat

You can install the package from this GitHub repository. You first need to install the devtools package.

install.packages("devtools")

Then install the NearRepeat package using the install_github function in the devtools package.

devtools::install_github("wsteenbeek/NearRepeat", force = TRUE)

Note that if you also want to view the vignettes directly within R, the command should read:

devtools::install_github("wsteenbeek/NearRepeat", build_vignettes = TRUE, force = TRUE)

After a few moments you should see “creating vignettes …” and at that point you have to wait a few minutes. Because of the wait time and because you may need additional packages to build vignettes, the default install_github() call does not build vignettes. Alternatively, the vignettes can also be viewed directly online (see section ‘Vignettes’).

Package NearRepeat needs packages future.apply and ggplot2 to function correctly. If these are not installed already, following the steps above should install these dependencies as well. If this doesn’t happen, it’s best to run install.packages(c("future.apply", "ggplot2")) and try the steps above again.

The package currently has one main function, also called NearRepeat(). See the helpfile to see examples of the function in action:

library(NearRepeat)
?NearRepeat

Vignettes are a long-form guide to an R package. Read these to get a relatively easy introduction to the package and its various options. You can view the available vignettes using:

browseVignettes("NearRepeat")

To directly read the vignettes rather than going through browseVignettes("NearRepeat") you can use:

vignette("NearRepeat", package = "NearRepeat")
vignette("NearRepeat_breaks", package = "NearRepeat")
vignette("NRC", package = "NearRepeat")
vignette("prepare_data", package = "NearRepeat")

Finally, the vignettes can be viewed directly, without installing the package, using the following links:

• NearRepeat: introduction to the package
• NearRepeat_breaks: a detailed description of how bandwidths can be specified
• NRC: a comparison to the Near Repeat Calculator
• prepare_data: how data was downloaded and prepped for the examples in the vignettes

This package is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License, version 3, as published by the Free Software Foundation.

This program is distributed in the hope that it will be useful, but without any warranty; without even the implied warranty of merchantability or fitness for a particular purpose. See the GNU General Public License for more details.

A copy of the GNU General Public License, version 3, is available at https://www.r-project.org/Licenses/GPL-3