inst/tinytest/test-modelplot.R

In modelsummary: Summary Tables and Plots for Statistical Models and Data: Beautiful, Customizable, and Publication-Ready

source("helpers.R")
if (ON_CI) exit_file("plot local only")
requiet("tinysnapshot")
using("tinysnapshot")
requiet("sandwich")

# single model
mod <- lm(hp ~ mpg + drat, data = mtcars)
p <- modelplot(mod)
expect_snapshot_plot(p, "modelplot-vanilla")

mod <- lm(hp ~ mpg + drat, data = mtcars)
p <- modelplot(mod, coef_omit = "Interc")
expect_snapshot_plot(p, "modelplot-coef_omit")

params <- list(
  ggplot2::geom_vline(xintercept = 0, color = "orange"),
  ggplot2::annotate(
    "rect",
    alpha = .1,
    xmin = -.5,
    xmax = .5,
    ymin = -Inf,
    ymax = Inf
  ),
  ggplot2::geom_point(
    ggplot2::aes(y = term, x = estimate),
    alpha = .3,
    size = 10,
    color = "red",
    shape = "square"
  )
)
p <- modelplot(
  mod,
  coef_map = c("drat" = "Rear axle ratio", "mpg" = "Miles / gallon"),
  color = "green",
  background = params
)
expect_snapshot_plot(p, "modelplot-coef_map_color_shape_background")

# multiple models
mod <- list(
  lm(hp ~ mpg + drat, data = mtcars),
  lm(hp ~ mpg, data = mtcars)
)
p <- modelplot(mod)
expect_snapshot_plot(p, "modelplot-multiple_plots_vanilla")

mod <- list(
  lm(hp ~ mpg + drat, data = mtcars),
  lm(hp ~ mpg, data = mtcars)
)
p <- modelplot(mod, facet = TRUE)
expect_snapshot_plot(p, "modeplot-multiple_plots_facet")

# preserve model order
mod <- list(
  "C" = lm(mpg ~ hp, data = mtcars),
  "A" = lm(mpg ~ hp + drat + vs, data = mtcars),
  "B" = lm(mpg ~ hp + drat, data = mtcars)
)
p <- modelplot(mod, draw = TRUE)
expect_snapshot_plot(p, "modeplot-model_order")

# conf_level=NULL
mod <- lm(hp ~ mpg + drat, data = mtcars)
p <- modelplot(mod, draw = FALSE, conf_level = NULL)
expect_inherits(p, "data.frame")
expect_equivalent(nrow(p), 3)

# vcov
mod <- list(
  lm(hp ~ mpg + drat, data = mtcars),
  lm(hp ~ mpg + drat, data = mtcars)
)

so <- list(vcov, sandwich::vcovHC)
p1 <- modelplot(mod, vcov = so, draw = FALSE)
p2 <- modelplot(mod, vcov = "HC3", draw = FALSE)

known <- c(
  165.179327669237,
  182.406373565931,
  -13.6502180401172,
  -15.0390897152001,
  -22.1832974370102,
  -28.1858724755655
)
expect_equivalent(p1$conf.low, known)

known <- c(
  182.406373565931,
  182.406373565931,
  -15.0390897152001,
  -15.0390897152001,
  -28.1858724755655,
  -28.1858724755655
)

known <- c(
  182.406373565931,
  182.406373565931,
  -15.0390897152001,
  -15.0390897152001,
  -28.1858724755655,
  -28.1858724755655
)
expect_equivalent(p2$conf.low, known)

# single model without ci
mod <- lm(hp ~ mpg + drat, data = mtcars)
p <- modelplot(mod, conf_level = NULL)

# multiple models without ci
mod <- list(
  lm(hp ~ mpg, data = mtcars),
  lm(hp ~ mpg + drat, data = mtcars),
  lm(hp ~ mpg + drat + am, data = mtcars)
)
p <- modelplot(mod, facet = TRUE, conf_level = NULL)
p <- modelplot(mod, facet = FALSE, conf_level = NULL)

# car::deltaMethod: Issue #491
requiet("car")
fit <- lm(mpg ~ disp + hp, mtcars)
fit_dm <- car::deltaMethod(fit, "disp / hp", rhs = 1)
p <- modelplot(fit_dm)

# Issue #529
requiet("survey")
dat <- mtcars
dat$weights <- dat$mpg / dat$disp
des <- svydesign(ids = ~1, data = dat, weights = dat$weights)
mod <- svyglm(
  vs ~ hp + drat + mpg + disp,
  design = des,
  family = quasibinomial()
)
p <- suppressWarnings(modelplot(mod, color = "red"))
expect_inherits(p, "gg")

# Issue #682
mod <- lm(mpg ~ hp, mtcars)
expect_silent(modelplot(mod, ci_method = "wald"))

# Issue #861: linewidth
mod <- lm(hp ~ mpg + drat, data = mtcars)
p <- modelplot(mod, linewidth = 3)
expect_snapshot_plot(p, "modelplot-issue861")