Plot methods for the package's results

Every result the package returns draws itself with plot(): no arguments beyond the object are needed, and every method takes the same customisation (main, col or palette, ... for the base-graphics call, the grid from the session option; see bricklayer_plot_options). Titles fit their panel: a title too wide for it is drawn a step smaller, and in a panel set one that still does not fit becomes a letter with the key printed under the panels (the returned data carry it as the panel_key attribute); the left margin is sized from the widest tick label, so an axis title never overlaps the numbers. (Session defaults: see bricklayer_plot_options for the session defaults). Each method returns the data it drew, invisibly, as a data frame.

Usage

# S3 method for class 'rmbl_hotspot'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_case_flow'
plot(
  x,
  ...,
  what = c("flow", "clearance"),
  main = NULL,
  col = NULL,
  palette = NULL
)

# S3 method for class 'rmbl_life_table'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_disposition_mix'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_density'
plot(x, ..., points = TRUE, main = NULL,
  col = NULL, palette = NULL)

# S3 method for class 'rmbl_knox'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_clearance_decomposition'
plot(x, ..., what = NULL,
  max_parts = 20L, main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_stock_flow'
plot(
  x,
  ...,
  what = c("adp", "alos", "people", "days"),
  main = NULL,
  col = NULL,
  palette = NULL
)

# S3 method for class 'rmbl_yoy'
plot(
  x,
  ...,
  what = c("pct_change", "value"),
  main = NULL,
  col = NULL,
  palette = NULL
)

# S3 method for class 'rmbl_rate'
plot(x, ..., period = NULL, main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_share'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_rate_change'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_band_sensitivity'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_drift_calibration'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_drift'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_benford'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_power'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_freq'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_missingness'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_region_coverage'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_outliers'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_correlations'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_cortable'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_falsification'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_bounds'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_crf'
plot(x, ..., exposure = NULL, main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_burden'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_equity'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_pollution_report'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_footprint'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_field'
plot(x, ..., contours = 8L, main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_particles'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_count_trend'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'bricklayer_analysis'
plot(x, ..., which = NULL, main = NULL)

# S3 method for class 'rmbl_svy_estimate'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

# S3 method for class 'rmbl_svy_compare'
plot(x, ..., main = NULL, col = NULL, palette = NULL)

Arguments

x

The result.

...

Passed to the base-graphics call (plot, barplot or image); for bricklayer_analysis, passed on to each panel's method.

main

Title; each method has a default.

col

Colours, overriding the palette.

palette

A bricklayer_palette name, overriding the session option.

what

For rmbl_stock_flow, the measures to draw (one panel each): any of "adp", "alos", "people", "days", "flow_rate", "stock_rate". For rmbl_yoy, "pct_change" or "value". For rmbl_case_flow, "flow" and/or "clearance"; for rmbl_clearance_decomposition, any of "rates", "mix" and "effects" (the default fits the device).

points

For rmbl_density, whether to draw the incidents over the surface.

max_parts

For rmbl_clearance_decomposition, the most parts drawn: past it, the parts that moved the aggregate most, and the rest summed into one row.

period

For rmbl_rate, a by column holding the period: the rate is drawn over time with its interval band, one panel per group of the other by columns.

exposure

For rmbl_crf and rmbl_burden, the exposure range to draw the curve over; the default runs from 0 to 1.5 times the exposure in the object.

contours

Number of contour lines over a field image (0 for none).

which

For bricklayer_analysis, the panels: any of "change", "rates", "rate_change"; NULL draws every panel the analysis has.

Value

The data drawn, invisibly.

Details

rmbl_stock_flow

(stock_flow) the average daily population, length of stay, people and person-days over the periods, one panel per measure in what; with an exposure, the flow and stock rates too.

rmbl_yoy

(yoy) the period-over-period percent change with its interval, one row per period (and group), coloured by verdict; what = "value" draws the values instead.

rmbl_rate, rmbl_share, rmbl_rate_change

(rate) the rate, share or percent change per group with its interval.

rmbl_band_sensitivity

(band_sensitivity) the statistic against the cap applied to the open band.

rmbl_drift_calibration, bricklayer_drift

(drift_calibrate, capsule_drift) the false-alarm rate per column against alpha; the population stability index per column with the drifted columns marked.

bricklayer_benford

(benford_test) observed first-digit proportions against Benford's law.

bricklayer_power

(capsule_power) the power curve: detection rate against injected effect size, with the 0.8 line.

bricklayer_freq, bricklayer_missingness, rmbl_region_coverage

(frequency_table, missingness_pattern, region_coverage) horizontal bars.

bricklayer_outliers

(mahalanobis_outliers) the distance of every row, outliers marked.

bricklayer_correlations, bricklayer_cortable

(top_correlations, correlation_table) the correlation of every pair as a diverging bar.

bricklayer_falsification

(capsule_falsify) the permutation null distribution with the observed statistic.

rmbl_bounds

(published_bounds) each value inside its published-rounding envelope.

rmbl_crf, rmbl_burden, rmbl_equity, rmbl_pollution_report

(crf_pm25, pollution_burden, exposure_concentration_index, verify_pollution) the concentration-response curve with the exposure marked; baseline against attributable cases; the concentration curve of exposure over income rank; and, for a report, all three on one page.

rmbl_footprint

(compute_footprint) the CPU, memory and total emission in one unit chosen to fit, with the car-kilometre and tree-month equivalents as a caption; "Too short to measure" when the work took no measurable time.

rmbl_field, rmbl_particles

(plume_field, advection_diffusion_2d, lagrangian_particles) a concentration field as an image with contours; particle positions, with the counting grid behind them when one was given.

rmbl_count_trend

(count_trend) the counts with the fitted trend.

bricklayer_analysis

(analyse_table) the change, rates and rate change of a published table, one panel each.

rmbl_case_flow

(case_flow) cases received and disposed, shaded where the inventory grew or fell, and the clearance rate with its interval and rolling value.

rmbl_life_table

(life_table) the share of cases still open at each band edge with its interval, and the median.

rmbl_disposition_mix

(disposition_mix) the shares of each outcome over the periods, stacked.

rmbl_density

(crime_density) the density of incidents as a heat map, with the highest-density regions outlined.

rmbl_knox

(knox_test) the Knox ratio of each distance-by-time cell, coloured on a log scale around 1.

rmbl_hotspot

(hotspot_persistence) the share of places holding half the incidents and the overlap of hot spots with the period before, over time, and a map of the cells coloured by how often they were hot.

rmbl_svy_estimate, rmbl_svy_compare

(survey_estimate, survey_compare) each domain's estimate with its interval, flagged domains marked; each difference from the reference domain with its interval.

rmbl_clearance_decomposition

(clearance_decomposition) the parts' rates in both comparisons with the aggregate on top, where the workload moved (coloured by whether a part clears above or below the average), and the change as a waterfall of rate and mix effects: the pictures that show a Simpson's-paradox reversal and why it happened.

Examples

sf <- stock_flow(days = c(13500, 14600, 15100), people = c(300, 320, 310),
                 period = 2021:2023)
plot(sf)

plot(sf, what = "alos", col = "firebrick", main = "Length of stay")

y <- yoy(data.frame(year = 2019:2023, n = c(400, 420, 390, 460, 450)), value = "n",
         period = "year")
plot(y)

# small counts: each percentage carries its exact interval
plot(yoy(data.frame(year = 2019:2023, n = c(10, 12, 11, 15, 14)), value = "n",
         period = "year"))

plot(benford_test(c(1, 12, 123, 1234, 2, 23, 3, 31, 4, 45, 5, 6, 7, 8, 9)))

plot(crf_no2(25))

plot(compute_footprint(for (i in 1:20) sum(sqrt(seq_len(2e5))), location = "CA",
                       cpu_power_w = 45, memory_gb = 8))

# the courts and crime data shipped with the package
ext <- function(f) {
  read.csv(system.file("extdata", f, package = "rmoriebricklayer"))
}
fl <- ext("ocj_flow_ontario.csv")
plot(case_flow(fl$received, fl$disposed, pending = fl$pending, period = fl$month))

td <- ext("ocj_time_to_disposition.csv")
open <- unlist(td[nrow(td), 2:5]) * fl$disposed[nrow(fl)]
plot(life_table(c(0, 10, 15, 18, Inf), open))

dm <- ext("ocj_dispositions.csv")
plot(disposition_mix(dm[, -(1:2)], period = dm$year))

og <- ext("ocj_offence_groups.csv")
og <- og[og$year %in% c(2014, 2022), ]
plot(clearance_decomposition(og$disposed, og$received, og$offence_group, og$year))

be <- ext("tps_break_enter_d14_2024.csv")
be <- be[!is.na(be$lon), ]
plot(crime_density(be$lon, be$lat, n_grid = 60))

# \donttest{
plot(knox_test(be$lon, be$lat, as.Date(be$occ_date), c(100, 200, 400), c(7, 14, 28),
               nsim = 99, seed = 1))

# }
plot(hotspot_persistence(be$lon, be$lat, quarters(as.Date(be$occ_date)), cell = 0.3))

# rates, shares and the change in a rate
rt <- data.frame(division = rep(c("n", "e", "w"), 2), year = rep(2022:2023, each = 3),
                 stops = c(120, 90, 60, 150, 85, 70),
                 residents = rep(c(1e4, 8e3, 5e3), 2))
r23 <- rt[rt$year == 2023, ]
plot(rate(r23, count = "stops", population = "residents", by = "division"))

plot(share(r23, count = "stops", by = "division"))

plot(rate_change(rt, count = "stops", population = "residents", period = "year",
                 by = "division"))

# a rate over time, one panel per group
plot(rate(rt, count = "stops", population = "residents", by = c("division", "year")),
     period = "year")

# complex survey estimates by domain, and the differences between domains
set.seed(3)
sv <- data.frame(region = rep(c("north", "south", "east"), length.out = 300),
                 smoker = rbinom(300, 1, 0.25), w = runif(300, 50, 150))
sd <- survey_design(sv, weights = "w", strata = "region")
plot(survey_estimate(sd, sv, "smoker", by = "region", stat = "proportion", levels = "1"))

plot(survey_compare(sd, sv, "smoker", by = "region", stat = "proportion", levels = "1"))

# a published table in one call: change, rates and rate change
tab <- data.frame(year = rep(2020:2023, each = 2), g = rep(c("a", "b"), 4),
                  n = c(10, 20, 12, 22, 11, 25, 15, 24), pop = rep(c(1000, 2000), 4))
plot(analyse_table(tab, value = "n", period = "year", by = "g", population = "pop"))

plot(count_trend(c(21000, 23000, 22000, 25000, 24000)))

# banded tables
plot(band_sensitivity(parse_bands(c("0-17", "18-24", "25-44", "45+")), c(10, 20, 30, 40),
                      caps = c(60, 70, 80, 90)))

plot(published_bounds(c(10, 20), rounding = 5))

plot(region_coverage(c("a", "b", "c"), c(100, 200, 300), c(1L, 0L, 2L)))

# drift between two extracts, and how often the test cries wolf
set.seed(1)
dd <- data.frame(a = rnorm(400), b = sample(letters[1:3], 400, TRUE))
plot(capsule_drift(dd, transform(dd, a = a + 3)))

plot(drift_calibrate(dd, n = 20L))

# describing a table
plot(frequency_table(data.frame(g = c("a", "b", "b", "c", "c", "c")), "g"))

plot(missingness_pattern(data.frame(a = c(1, NA, 3, NA), b = c(NA, NA, 1, 2))))

hw <- data.frame(h = rnorm(100), w = rnorm(100))
hw$h[1] <- 15
plot(mahalanobis_outliers(hw))

sq <- data.frame(a = 1:20, b = (1:20)^2, c = rev(1:20))
plot(top_correlations(sq))

plot(correlation_table(sq))

# power and falsification of a statistic
xy <- data.frame(x = rnorm(60), y = rnorm(60))
shift <- function(z, s) {
  z$y <- z$y + s * z$x
  z
}
plot(capsule_power(xy, function(z) cor(z$x, z$y), inject = shift, treatment = "x",
                   sizes = c(0, 0.25, 0.5), n = 19, reps = 3, seed = 1))

plot(capsule_falsify(xy, function(z) cor(z$x, z$y), treatment = "x", n = 29, seed = 1))

# air pollution: burden, equity, and the full report
plot(pollution_burden(25, 1, 0.008, 1e6, pollutant = "NO2"))

plot(exposure_concentration_index(data.frame(exposure = c(4, 1, 8, 2),
                                             income = c(2, 1, 4, 3)),
                                  "exposure", "income"))

plot(verify_pollution("no2", demo = TRUE))

plot(lagrangian_particles(200, 0, 0, 1, 0.5, 1, 1, 1, 20,
                          grid = c(-10, 40, -15, 25, 10, 8)))

plot(plume_field(q = 100, u = 5, h = 50, x = seq(100, 2000, by = 100),
                 y = seq(-300, 300, by = 50)))