Ranks units (neighbourhoods, divisions) on several kinds of crime at
once: each category's counts are standardised across the units (minus
the mean, over the standard deviation) and the standardised values are
summed with weights. A category whose counts do not vary contributes 0
to every unit. A unit with no row for a category counts 0 there, and
its standardised value is computed with that 0.
Usage
zcomp(x, unit, cat, w = NULL)
Arguments
- x
The counts: numbers from 0, one per row.
- unit
The unit of each row: labels, at least two units.
- cat
The category of each row: labels.
- w
NULL for equal weights, or the weight of each category: numbers
from 0, named by category or in the order of the categories.
Value
An object of class zcomp, a list: tab, a data frame with one
row per unit (unit, comp the composite, rank with 1 the highest
and ties sharing the lowest rank); cat, one row per category (cat,
mean, sd and w); z, the matrix of standardised counts, units by
categories.
Examples
d <- data.frame(hood = rep(c("a", "b", "c"), 2),
kind = rep(c("assault", "theft"), each = 3),
n = c(5, 9, 1, 30, 22, 41))
zcomp(d$n, d$hood, d$kind)
#> zcomp: composite of 2 standardised categories
#> unit comp rank
#> a -0.10482848 3
#> b 0.05654365 1
#> c 0.04828484 2
