Plot Posterior Distributions for btdFoot Objects

Description

Plots for the posterior distributions of team log-strengths and other parameters with customizable plot types and facets.

Usage

plot_btdPosterior(
  x,
  pars = "logStrength",
  plot_type = "boxplot",
  teams = NULL,
  ncol = NULL,
  scales = NULL
)

Arguments

x	An object of class btdFoot.
pars	A character string specifying the parameter to plot. Choices are "logStrength", "logTie", and "home". Default is "logStrength".
plot_type	A character string specifying the type of plot. Choices are "boxplot" and "density". Default is "boxplot".
teams	An optional character vector specifying team names to include in the posterior boxplots or density plots. If NULL, all teams are included.
ncol	An optional integer specifying the number of columns in the facet wrap when using a dynamic Bayesian Bradley-Terry-Davidson model. Default is 8.
scales	An optional character string specifying the scales for the facets when using a dynamic Bayesian Bradley-Terry-Davidson model. Options include "free", "fixed", "free_x", and "free_y". Default is "free_x".

Value

A ggplot object displaying:

• For pars="logStrength":

  • Dynamic BTD: Faceted boxplots or density plots (including the 95% credible interval) of posterior log-strengths by team and period.

  • Static BTD: Boxplots or density plots (including the 95% credible interval) of posterior log-strengths for each team.

• For pars="logTie" or pars="home": A single boxplot or density plot with 95% credible interval.

Author(s)

Roberto Macrì Demartino roberto.macridemartino@deams.units.it.

Examples

## Not run: 
if (instantiate::stan_cmdstan_exists()) {
  library(dplyr)

  # Load example data
  data("italy")

  # Prepare the data
  italy_2020_2021_rank <- italy %>%
    select(Season, home, visitor, hgoal, vgoal) %>%
    filter(Season %in% c("2020", "2021")) %>%
    mutate(match_outcome = case_when(
      hgoal > vgoal ~ 1, # Home team wins
      hgoal == vgoal ~ 2, # Draw
      hgoal < vgoal ~ 3 # Away team wins
    )) %>%
    mutate(periods = case_when(
      row_number() <= 190 ~ 1,
      row_number() <= 380 ~ 2,
      row_number() <= 570 ~ 3,
      TRUE ~ 4
    )) %>% # Assign periods based on match number
    select(periods,
      home_team = home,
      away_team = visitor, match_outcome
    )

  # Fit the Bayesian Bradley-Terry-Davidson model with dynamic ranking
  fit_rank_dyn <- btd_foot(
    data = italy_2020_2021_rank,
    dynamic_rank = TRUE,
    rank_measure = "median",
    iter_sampling = 1000,
    parallel_chains = 2,
    chains = 2
  )

  # Plot posterior distributions with default settings
  plot_btdPosterior(fit_rank_dyn)

  # Plot posterior distributions for specific teams with customized facets
  plot_btdPosterior(
    fit_rank_dyn,
    teams = c("AC Milan", "AS Roma", "Juventus", "Inter"),
    ncol = 2
  )

  plot_btdPosterior(
    fit_rank_dyn,
    plot_type = "density",
    teams = c("AC Milan", "AS Roma", "Juventus", "Inter"),
    ncol = 2
  )
}

## End(Not run)