### Install bayesplot Package

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Install the stable version from CRAN or the development version from GitHub.

```r
install.packages("bayesplot")
```

```r
devtools::install_github("stan-dev/bayesplot", dependencies = TRUE, build_vignettes = FALSE)
```

--------------------------------

### Install bayesplot Development Version from GitHub

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Install the latest development version of bayesplot directly from GitHub. This requires the devtools package and may not include vignettes by default.

```r
if (!require("devtools")) {
  install.packages("devtools")
}
devtools::install_github("stan-dev/bayesplot", dependencies = TRUE, build_vignettes = FALSE)
```

--------------------------------

### Install bayesplot from CRAN

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Use this command to install the stable version of the bayesplot package from the Comprehensive R Archive Network (CRAN).

```r
install.packages("bayesplot")
```

--------------------------------

### Extract data for PPC plots

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

New functions ending in `_data` return data prepared for plotting, without generating the plot itself. Examples include `ppc_intervals_data()`, `ppc_ribbon_data()`, `mcmc_parcoord_data()`, `mcmc_rhat_data()`, and `mcmc_neff_data()`.

```R
ppc_intervals_data(y, y_rep)
ppc_ribbon_data(y, y_rep)
mcmc_parcoord_data(fit)
mcmc_rhat_data(fit)
mcmc_neff_data(fit)
```

--------------------------------

### Get data for PPC plots

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

The `ppc_data()` function returns the data prepared for plotting by many PPC functions, useful for custom visualizations.

```R
ppc_data(y, y_rep, train_indices = "all")
```

--------------------------------

### Grouped Posterior Predictive Checks with bayesplot

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Performs grouped posterior predictive checks using ppc_stat_grouped, calculating a specified statistic (e.g., median) for the observed data and simulated data within defined groups. This example demonstrates piping for a more concise workflow.

```r
# also works nicely with piping
library("dplyr")
color_scheme_set("brightblue")
fit %>%
  posterior_predict(draws = 500) %>%
  ppc_stat_grouped(y = mtcars$mpg, 
                   group = mtcars$carb, 
                   stat = "median")
```

--------------------------------

### Overlay Posterior Predictive Densities with bayesplot

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Visualizes the distribution of actual data against simulated data from the posterior predictive distribution using ppc_dens_overlay. Requires setting a color scheme and obtaining posterior predictions.

```r
color_scheme_set("red")
ppc_dens_overlay(y = fit$y, 
                 yrep = posterior_predict(fit, draws = 50))
```

--------------------------------

### MCMC Intervals Colored by R-hat

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Illustrates coloring `mcmc_intervals()` plots by R-hat values. This helps identify parameters with potential convergence issues.

```R
fake_rhat <- c(1.0, 1.07, 1.3, 1.01, 1.15, 1.005)
color_scheme_set("blue")
mcmc_intervals(x, rhat = fake_rhat)
```

--------------------------------

### Visualize posterior density with ridgelines

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

New functions `mcmc_dens_chains()` and `mcmc_areas_ridges()` visualize posterior density using ridgelines, suitable for hierarchical models.

```R
mcmc_dens_chains(fit, pars = "beta[1]")
mcmc_areas_ridges(fit, pars = "beta[1]")
```

--------------------------------

### Test Package as CRAN

Source: https://github.com/stan-dev/bayesplot/wiki/bayesplot-development-notes

Set environment variable to simulate CRAN testing conditions and run package tests.

```R
Sys.setenv(NOT_CRAN = "false")
devtools::test()
```

--------------------------------

### LOO PIT predictive check

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

The `ppc_loo_pit_overlay()` function provides an improved visualization for LOO PIT predictive checks.

```R
ppc_loo_pit_overlay(y, y_rep)
```

--------------------------------

### Handle discrete data in ppc_ecdf_overlay

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

The `discrete` argument in `ppc_ecdf_overlay()` can be set to TRUE to make the geom more appropriate for discrete data.

```R
ppc_ecdf_overlay(y, y_rep, discrete = TRUE)
```

--------------------------------

### MCMC Histograms by Chain

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Creates per-chain histograms using a facet grid. Useful for visualizing the distribution of draws for a specific parameter across different chains.

```R
color_scheme_set("orange")
mcmc_hist_by_chain(x, regex_pars = "beta")
```

--------------------------------

### MCMC Interval Plots

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Visualizes quantile-based credible intervals using `mcmc_intervals()`. Supports customization of interval widths, point estimates, and coloring by R-hat.

```R
library(bayesplot)
library(ggplot2)
x <- example_mcmc_draws(params = 6)

color_scheme_set("brightblue")
mcmc_intervals(x)
```

```R
mcmc_intervals(x, pars = c("beta[1]", "beta[2]"), prob = 0.8, prob_outer = 0.95,
               point_est = "mean")
```

--------------------------------

### MCMC Trace Plots

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Generates standard or highlight-chain trace plots of MCMC draws over iterations. Useful for visualizing MCMC convergence and behavior.

```APIDOC
## MCMC Trace Plots

### `mcmc_trace()` / `mcmc_trace_highlight()`

Standard or highlight-chain trace plots of MCMC draws over iterations. Accepts a 3-D array `[iterations × chains × parameters]`, a list of matrices, or a data frame. The `n_warmup` argument shades warmup iterations gray. NUTS divergences can be annotated as rug marks via the `np` argument (from `nuts_params()`).

```r
library(bayesplot)
library(ggplot2)

x <- example_mcmc_draws(chains = 4, params = 6)

# Basic trace plot for parameters matching "beta"
color_scheme_set("viridis")
mcmc_trace(x, regex_pars = "beta")

# Zoom to iterations 100–130, shade warmup, annotate divergences
# (with a real Stan fit, pass nuts_params(fit) to 'np')
color_scheme_set("mix-blue-red")
mcmc_trace(
  x,
  pars   = c("alpha", "sigma"),
  n_warmup = 100,
  window = c(100, 130),
  size   = 0.8
) + 
  panel_bg(fill = "gray95", color = NA) + 
  legend_move("top")

# Highlight a single chain (chain 2 is opaque; others are faded)
color_scheme_set("brightblue")
mcmc_trace_highlight(x, pars = "sigma", highlight = 2, size = 2)
```
```

--------------------------------

### Compare Observed Data vs. Posterior Predictive Distributions

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Use these functions to compare the distribution of observed data `y` against replications `yrep` drawn from the posterior predictive distribution. `yrep` should be a matrix with one row per posterior draw and one column per observation. Ensure `bayesplot` and `ggplot2` are loaded.

```r
library(bayesplot)
library(ggplot2)

y    <- example_y_data()          # observed outcome vector (length N)
yrep <- example_yrep_draws()      # S × N matrix of posterior predictive draws
group <- example_group_data()

# Overlaid density curves: 25 yrep draws (light) + y (dark)
color_scheme_set("brightblue")
ppc_dens_overlay(y, yrep[1:25, ])

# ECDF overlay: step functions for yrep, darker step for y
ppc_ecdf_overlay(y, yrep[sample(nrow(yrep), 25), ])

# Histograms: one facet per yrep draw + one for y
color_scheme_set("red")
ppc_hist(y, yrep[1:8, ])

# Box plots with notches
ppc_boxplot(y, yrep[1:8, ])

# Density and ECDF overlays faceted by group
color_scheme_set("teal")
ppc_dens_overlay_grouped(y, yrep[1:25, ], group = group)
ppc_ecdf_overlay_grouped(y, yrep[1:25, ], group = group)

# Violin grouped: yrep as violin, y overlaid as outline or points
color_scheme_set("gray")
ppc_violin_grouped(y, yrep, group, size = 1.5)
ppc_violin_grouped(y, yrep, group, alpha = 0, y_draw = "both",
                   y_size = 1.5, y_alpha = 0.5, y_jitter = 0.33)

# PIT-ECDF: probability integral transform check with 99% bands
ppc_pit_ecdf(y, yrep, prob = 0.99, plot_diff = FALSE)
ppc_pit_ecdf(y, yrep, prob = 0.99, plot_diff = TRUE)   # difference plot
```

--------------------------------

### MCMC Effective Sample Size Ratio Histogram

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Displays the effective sample size (N_eff) to total sample size (N) ratio as a histogram using `mcmc_neff_hist()`. Summarizes the distribution of these ratios.

```R
mcmc_neff_hist(neff_vals)
```

--------------------------------

### MCMC Distribution Plots

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Visualizes marginal posterior distributions of MCMC draws using histograms, density plots, violin plots, and quantile dot plots.

```APIDOC
## MCMC Distribution Plots

### `mcmc_hist()` / `mcmc_dens()` / `mcmc_dens_overlay()` / `mcmc_violin()` / `mcmc_dots()`

Marginal distribution summaries of posterior draws with all chains merged or separated. `mcmc_dens_overlay()` draws one density curve per chain; `mcmc_dens_chains()` uses ridgelines. `mcmc_violin()` shows chain-specific violins. `mcmc_dots()` / `mcmc_dots_by_chain()` produce quantile dot plots (requires **ggdist**).

```r
library(bayesplot)
x <- example_mcmc_draws(chains = 4, params = 6)


```

--------------------------------

### Plot Posterior Distributions with bayesplot

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Generates a plot of posterior distributions for specified parameters using mcmc_areas. Requires loading bayesplot, rstanarm, and ggplot2 libraries, and fitting a model with stan_glm.

```r
library("bayesplot")
library("rstanarm")
library("ggplot2")

fit <- stan_glm(mpg ~ ., data = mtcars)
posterior <- as.matrix(fit)

plot_title <- ggtitle("Posterior distributions",
                      "with medians and 80% intervals")
mcmc_areas(posterior, 
           pars = c("cyl", "drat", "am", "wt"), 
           prob = 0.8) + plot_title
```

--------------------------------

### MCMC Histograms

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Generates histograms of MCMC draws for parameters. Supports free x-axis scales per facet and named transformation strings for labeling axes.

```R
color_scheme_set("brightblue")
mcmc_hist(x)
```

```R
mcmc_hist(
  x,
  pars           = c("beta[1]", "beta[2]", "sigma"),
  transformations = list(sigma = "log")
)
```

--------------------------------

### MCMC Density Overlay

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Plots overlaid per-chain density curves. Allows for comparison of density estimates across chains for selected parameters.

```R
color_scheme_set("mix-teal-pink")
mcmc_dens_overlay(x, pars = c("sigma", "beta[2]"),
                  facet_args = list(nrow = 2))
```

--------------------------------

### Configure Knitr Chunk Options

Source: https://github.com/stan-dev/bayesplot/blob/master/vignettes/children/SETTINGS-knitr.txt

Sets default options for all R code chunks processed by Knitr. This ensures consistent figure generation for bayesplot visualizations.

```r
stopifnot(require("knitr"))
library("bayesplot")
knitr::opts_chunk$set(
  dev = "png",
  dpi = 150,
  fig.asp = 0.618,
  fig.width = 5,
  out.width = "60%",
  fig.align = "center",
  comment = NA,
  eval = if (isTRUE(exists("params"))) params$EVAL else FALSE
)
```

--------------------------------

### MCMC Trace Plot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Generate standard or highlight-chain trace plots for MCMC draws. The n_warmup argument shades warmup iterations, and the np argument can annotate divergences. Plots can be customized with ggplot2 functions.

```r
library(bayesplot)
library(ggplot2)

x <- example_mcmc_draws(chains = 4, params = 6)

# Basic trace plot for parameters matching "beta"
color_scheme_set("viridis")
mcmc_trace(x, regex_pars = "beta")

# Zoom to iterations 100–130, shade warmup, annotate divergences
# (with a real Stan fit, pass nuts_params(fit) to 'np')
color_scheme_set("mix-blue-red")
mcmc_trace(
  x,
  pars   = c("alpha", "sigma"),
  n_warmup = 100,
  window = c(100, 130),
  size   = 0.8
) +
  panel_bg(fill = "gray95", color = NA) +
  legend_move("top")

# Highlight a single chain (chain 2 is opaque; others are faded)
color_scheme_set("brightblue")
mcmc_trace_highlight(x, pars = "sigma", highlight = 2, size = 2)
```

--------------------------------

### MCMC Trace Plot with RStan

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Generates an MCMC trace plot for specified parameters ('mu', 'tau') from an RStan fit. Requires the 'rstan' library and a fitted model object. Customizes plot appearance with color schemes and facet labels.

```r
library("rstan")
fit2 <- stan_demo("eight_schools", warmup = 300, iter = 700)
posterior2 <- extract(fit2, inc_warmup = TRUE, permuted = FALSE)

color_scheme_set("mix-blue-pink")
p <- mcmc_trace(posterior2,  pars = c("mu", "tau"), n_warmup = 300,
                facet_args = list(nrow = 2, labeller = label_parsed))
p + facet_text(size = 15)
```

--------------------------------

### MCMC Violin Plot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Creates violin plots for MCMC draws, showing the distribution of draws for each chain. Includes options for adding panel backgrounds and styling.

```R
color_scheme_set("green")
mcmc_violin(x) + panel_bg(color = "gray20", linewidth = 2, fill = "gray30")
```

--------------------------------

### Compare Posterior Predictive Test Statistics to Observed Values

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Use these functions to display the distribution of a test statistic `T(yrep)` across posterior predictive replications and compare it to the observed value `T(y)`. Any scalar-valued function can be used as the statistic. Ensure `bayesplot` is loaded.

```r
library(bayesplot)

y    <- example_y_data()
yrep <- example_yrep_draws()
group <- example_group_data()

# Distribution of the median across yrep; vertical line at median(y)
color_scheme_set("brightblue")
ppc_stat(y, yrep, stat = "median")
ppc_stat(y, yrep, stat = "sd") + legend_none()

# Custom statistic: 25th percentile
q25 <- function(y) quantile(y, 0.25)
ppc_stat(y, yrep, stat = "q25")

# Grouped test statistic — one facet per group level
color_scheme_set("teal")
ppc_stat_grouped(y, yrep, group, stat = "median")
ppc_stat_grouped(y, yrep, group, stat = "mad") + yaxis_text()

# 2D scatterplot of two test statistics across yrep
ppc_stat_2d(y, yrep, stat = c("mean", "sd"))

# Discrete data example
set.seed(42)
y_disc   <- rbinom(30, size = 1, prob = 0.3)
yrep_disc <- matrix(rbinom(3000, size = 1, prob = 0.5), nrow = 100)
ppc_stat(y_disc, yrep_disc, stat = "mean", discrete = TRUE)
```

--------------------------------

### Return data from mcmc_intervals and mcmc_areas

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

New functions `mcmc_intervals_data()` and `mcmc_areas_data()` return the data plotted by `mcmc_intervals()` and `mcmc_areas()`, respectively.

```R
mcmc_intervals_data(fit)
mcmc_areas_data(fit)
```

--------------------------------

### MCMC R-hat Lollipop Plot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Visualizes R-hat values using `mcmc_rhat()`. Parameters are plotted with lollipop segments, colored by R-hat quality (light/mid/dark).

```R
library(bayesplot)

rhat_vals <- c(runif(80, 1.0, 1.05), runif(15, 1.05, 1.1), runif(5, 1.1, 1.3))
names(rhat_vals) <- paste0("param_", seq_along(rhat_vals))

color_scheme_set("blue")
mcmc_rhat(rhat_vals)
```

--------------------------------

### MCMC R-hat Histogram

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Displays R-hat values as a histogram using `mcmc_rhat_hist()`. Provides a summary view of the distribution of R-hat statistics.

```R
mcmc_rhat_hist(rhat_vals)
```

--------------------------------

### MCMC Dot Plot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Generates a quantile dot plot, which requires the ggdist package. Useful for visualizing quantiles of posterior distributions for selected parameters.

```R
color_scheme_set("pink")
mcmc_dots(x, pars = c("alpha", "beta[2]"))
```

--------------------------------

### MCMC Scatterplot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Creates scatterplots of two parameters from MCMC draws using `mcmc_scatter()`. Supports highlighting divergences and customizing plot aesthetics.

```R
library(bayesplot)
library(ggplot2)
x <- example_mcmc_draws(chains = 4, params = 6)

color_scheme_set("darkgray")
mcmc_scatter(
  as.matrix(x),
  pars     = c("alpha", "beta[1]"),
  size     = 1,
  alpha    = 0.5
)
```

--------------------------------

### MCMC Scatter Plot with Divergences

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Creates an MCMC scatter plot showing the relationship between two parameters ('tau', 'theta[1]') and highlighting divergent transitions. Uses 'nuts_params' to extract NUTS parameters and customizes the style for divergent points.

```r
color_scheme_set("darkgray")
mcmc_scatter(
  as.matrix(fit2),
  pars = c("tau", "theta[1]"), 
  np = nuts_params(fit2), 
  np_style = scatter_style_np(div_color = "green", div_alpha = 0.8)
)
```

--------------------------------

### Control faceting in grouped PPC plots

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

The `facet_args` argument in `ppc_stat_grouped()` and `ppc_stat_freqpoly_grouped()` allows control over ggplot2 faceting.

```R
ppc_stat_grouped(y, y_rep, group = group_var, facet_args = list(nrow = 2))
```

--------------------------------

### MCMC Pairwise Scatterplot Matrix

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Generates a pairwise scatterplot matrix using `mcmc_pairs()`. Useful for visualizing relationships between multiple parameters.

```R
color_scheme_set("blue")
mcmc_pairs(
  x,
  pars         = c("alpha", "sigma", "beta[1]"),
  off_diag_args = list(size = 0.5, alpha = 0.4)
)
```

--------------------------------

### NUTS Acceptance Rate Plot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Visualizes the NUTS (No-U-Turn Sampler) acceptance rate. Requires `nuts_params()` output from a Stan fit.

```R
library(bayesplot)
library(ggplot2)
# Requires a Stan fit; shown here with rstanarm
# library(rstanarm)
```

--------------------------------

### MCMC Distribution Plots

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Summarize marginal posterior distributions of MCMC draws. Functions include histograms (mcmc_hist), density plots (mcmc_dens, mcmc_dens_overlay, mcmc_dens_chains), violin plots (mcmc_violin), and quantile dot plots (mcmc_dots, mcmc_dots_by_chain).

```r
library(bayesplot)
x <- example_mcmc_draws(chains = 4, params = 6)
```

--------------------------------

### Theme Management

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Functions to manage the ggplot2 theme applied to bayesplot plots. Includes setting defaults, updating elements, and restoring original themes.

```APIDOC
## Theme Management Functions

### `theme_default()` / `bayesplot_theme_set()` / `bayesplot_theme_update()`

**Description:**
Manage the ggplot2 theme applied to all bayesplot plots. The default is `theme_default()`, a clean bw-based theme. Theme changes affect only bayesplots, not other ggplot objects.

### Usage:
```r
# Set to built-in default theme
bayesplot_theme_set()

# Set with custom ggplot2 theme
bayesplot_theme_set(theme_default(base_size = 9, base_family = "sans"))
bayesplot_theme_set(theme_classic())

# Update individual theme elements
bayesplot_theme_update(
  plot.title = element_text(size = 16, face = "bold"),
  panel.background = element_rect(fill = "gray98")
)

# Restore the default theme
bayesplot_theme_set()
```
```

--------------------------------

### NUTS Energy Diagnostic Plot

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Visualizes the NUTS energy diagnostic using the 'mcmc_nuts_energy' function. This plot helps assess the efficiency of the Hamiltonian Monte Carlo sampler. Requires NUTS parameters extracted from a fitted model.

```r
color_scheme_set("red")
np <- nuts_params(fit2)
mcmc_nuts_energy(np) + ggtitle("NUTS Energy Diagnostic")
```

--------------------------------

### MCMC Effective Sample Size Ratio Lollipop Plot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Visualizes the effective sample size (N_eff) to total sample size (N) ratio using `mcmc_neff()`. Parameters are plotted with lollipop segments, colored by ratio (high/ok/low).

```R
neff_vals <- c(runif(70, 0.5, 1.0), runif(20, 0.1, 0.5), runif(10, 0.0, 0.1))
color_scheme_set("purple")
mcmc_neff(neff_vals, size = 3)
```

--------------------------------

### MCMC Rank Plots

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Provides rank-based chain-mixing diagnostics, including rank histograms, overlaid rank step-lines, and empirical cumulative distribution functions (ECDFs).

```APIDOC
## MCMC Rank Plots

### `mcmc_rank_hist()` / `mcmc_rank_overlay()` / `mcmc_rank_ecdf()`

Rank-based chain-mixing diagnostics. Instead of raw values over iterations, ranks of the combined draws are histogrammed per chain. A uniform distribution across chains indicates good mixing. `mcmc_rank_ecdf()` adds simultaneous confidence bands.

```r
x <- example_mcmc_draws(chains = 4, params = 6)

color_scheme_set("viridisE")

# Rank histogram — one panel per (parameter × chain)
mcmc_rank_hist(x, pars = c("alpha", "sigma"), ref_line = TRUE)

# Overlaid rank step-lines — all chains on one panel per parameter
mcmc_rank_overlay(x, regex_pars = "beta", ref_line = TRUE)

# Rank ECDF with 99% simultaneous confidence bands
mcmc_rank_ecdf(x, prob = 0.99)

# ECDF difference plot: deviations from the ideal uniform distribution
mcmc_rank_ecdf(x, prob = 0.99, plot_diff = TRUE)
```
```

--------------------------------

### Posterior Predictive Intervals with RStanarm

Source: https://github.com/stan-dev/bayesplot/blob/master/README.md

Generates posterior predictive intervals for a regression model fitted with 'stan_glmer' from the 'rstanarm' package. Visualizes observed data against predicted intervals, with custom panel backgrounds and grid lines.

```r
color_scheme_set("purple")

fit <- stan_glmer(mpg ~ wt + (1|cyl), data = mtcars)
ppc_intervals(
  y = mtcars$mpg,
  yrep = posterior_predict(fit),
  x = mtcars$wt,
  prob = 0.5
) +
  labs(
    x = "Weight (1000 lbs)",
    y = "MPG",
    title = "50% posterior predictive intervals \nvs observed miles per gallon",
    subtitle = "by vehicle weight"
  ) +
  panel_bg(fill = "gray95", color = NA) +
  grid_lines(color = "white")
```

--------------------------------

### Set and View Color Schemes

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Manage the global color palette for bayesplot functions. Supports preset schemes, custom hex values, and mixed schemes. Use color_scheme_view() to visualize the current or multiple schemes.

```r
library(bayesplot)
library(ggplot2)

# Set a preset scheme
color_scheme_set("viridis")

# View the current scheme as a color swatch
color_scheme_view()

# Retrieve specific colors (indices 1–6 = lightest to darkest)
color_scheme_get(i = c(1, 3, 6))

# Compare multiple schemes side-by-side
color_scheme_view(c("pink", "teal", "viridisA"))

# Custom six-color scheme (light → dark)
orange_scheme <- c("#ffebcc", "#ffcc80", "#ffad33", "#e68a00", "#995c00", "#663d00")
color_scheme_set(orange_scheme)
color_scheme_view()

# Mix two preset schemes
color_scheme_set("mix-blue-pink")
x <- example_mcmc_draws(chains = 4, params = 6)
mcmc_hist(x, regex_pars = "beta")
```

--------------------------------

### Color Scheme Management

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Functions to set, retrieve, and visualize the global color palette used by bayesplot functions. Supports preset schemes and custom color definitions.

```APIDOC
## Color Scheme Management

### `color_scheme_set()` / `color_scheme_get()` / `color_scheme_view()`

Set, retrieve, or visualize the global color palette used by all bayesplot functions. Preset schemes include `"blue"`, `"brightblue"`, `"red"`, `"green"`, `"teal"`, `"pink"`, `"purple"`, `"gray"`, `"darkgray"`, `"yellow"`, `"viridis"` (and variants A–E), mix combos (`"mix-teal-pink"`), and ColorBrewer palettes (`"brewer-Blues"`). A custom scheme is specified as exactly six hex/color values from lightest to darkest.

```r
library(bayesplot)
library(ggplot2)

# Set a preset scheme
color_scheme_set("viridis")

# View the current scheme as a color swatch
color_scheme_view()

# Retrieve specific colors (indices 1–6 = lightest to darkest)
color_scheme_get(i = c(1, 3, 6))

# Compare multiple schemes side-by-side
color_scheme_view(c("pink", "teal", "viridisA"))

# Custom six-color scheme (light → dark)
orange_scheme <- c("#ffebcc", "#ffcc80", "#ffad33", "#e68a00", "#995c00", "#663d00")
color_scheme_set(orange_scheme)
color_scheme_view()

# Mix two preset schemes
color_scheme_set("mix-blue-pink")
x <- example_mcmc_draws(chains = 4, params = 6)
mcmc_hist(x, regex_pars = "beta")
```
```

--------------------------------

### MCMC Ridgeline Densities

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Generates ridgeline plots where parameters are on the y-axis and chains are represented as overlapping lines. Provides a compact display of density distributions.

```R
color_scheme_set("purple")
mcmc_dens_chains(x, pars = c("beta[1]", "beta[2]", "beta[3]"))
```

--------------------------------

### MCMC Rank Plots

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Visualize MCMC chain-mixing diagnostics using rank-based plots. mcmc_rank_hist() shows histograms of ranks per chain, mcmc_rank_overlay() plots overlaid rank step-lines, and mcmc_rank_ecdf() adds confidence bands or difference plots.

```r
x <- example_mcmc_draws(chains = 4, params = 6)

color_scheme_set("viridisE")

# Rank histogram — one panel per (parameter × chain)
mcmc_rank_hist(x, pars = c("alpha", "sigma"), ref_line = TRUE)

# Overlaid rank step-lines — all chains on one panel per parameter
mcmc_rank_overlay(x, regex_pars = "beta", ref_line = TRUE)

# Rank ECDF with 99% simultaneous confidence bands
mcmc_rank_ecdf(x, prob = 0.99)

# ECDF difference plot: deviations from the ideal uniform distribution
mcmc_rank_ecdf(x, prob = 0.99, plot_diff = TRUE)
```

--------------------------------

### Highlight divergences in mcmc_scatter

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

The `np` argument in `mcmc_scatter()` allows specifying NUTS parameters to highlight divergences in the plot.

```R
mcmc_scatter(fit, pars = c("beta[1]", "beta[2]"), np = nuts_params(fit, "divergent__"))
```

--------------------------------

### Arrange Multiple Bayespots in a Grid

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Arrange multiple bayesplots (ggplot objects) in a shared grid, optionally aligning x/y axes and adding titles. This function wraps `gridExtra::arrangeGrob()`.

```r
library(bayesplot)

y    <- example_y_data()
yrep <- example_yrep_draws()
stats <- c("sd", "median", "max", "min")

color_scheme_set("pink")
bayesplot_grid(
  plots     = lapply(stats, function(s) ppc_stat(y, yrep, stat = s)),
  titles    = stats,
  legends   = FALSE,
  grid_args = list(ncol = 2)
)
```

```r
# Share x-axis limits across multiple interval plots
x <- example_mcmc_draws(params = 6)
color_scheme_set("teal")
bayesplot_grid(
  mcmc_intervals(x, pars = c("alpha", "sigma")),
  mcmc_areas(x, regex_pars = "beta\[[1-3]\]"),
  xlim = c(-3, 3),
  grid_args = list(nrow = 1)
)
```

--------------------------------

### Retrieve Underlying Tidy Data Frames

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Companion `*_data()` functions return tidy data frames used by plot functions, enabling custom ggplot2 visualizations. These functions are useful for users who need fine-grained control over plot construction.

```r
library(bayesplot)
library(ggplot2)

x    <- example_mcmc_draws(chains = 4, params = 6)
y    <- example_y_data()
yrep <- example_yrep_draws()

# Retrieve trace data (includes value_rank for rank plots)
trace_df <- mcmc_trace_data(x, pars = c("alpha", "sigma"))
head(trace_df)
# parameter  value  value_rank  chain  iteration  ...
```

```r
# Retrieve interval summary data
int_df <- mcmc_intervals_data(x, prob = 0.5, prob_outer = 0.9)
# columns: parameter, ll, l, m, h, hh
```

```r
# Custom interval plot built from the data backend
ggplot(int_df, aes(y = parameter)) +
  geom_segment(aes(x = ll, xend = hh, yend = parameter), color = "gray60") +
  geom_segment(aes(x = l,  xend = h,  yend = parameter), linewidth = 2) +
  geom_point(aes(x = m), size = 3, shape = 21, fill = "white") +
  theme_bw() +
  labs(title = "Custom interval plot from mcmc_intervals_data()")
```

```r
# Retrieve PPC data for custom plotting
ppc_df <- ppc_data(y, yrep[1:10, ])
# columns: y_id, rep_id, rep_label, value, is_y, is_y_label
```

```r
# Retrieve test-statistic data
stat_df <- ppc_stat_data(y, yrep, stat = "median")
# columns: variable (y vs yrep), value of stat
```

--------------------------------

### Plot Posterior Predictive Intervals for Observations

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Use these functions to plot posterior predictive intervals for each observation, sorted or arranged along an x-axis variable, and overlaid with the observed values. This is useful for detecting systematic model misfits. Ensure `bayesplot` and `ggplot2` are loaded.

```r
library(bayesplot)
library(ggplot2)

y    <- example_y_data()
yrep <- example_yrep_draws()
x    <- example_x_data()        # covariate for x-axis ordering
group <- example_group_data()

# 50% inner and 90% outer predictive intervals vs observed y
color_scheme_set("brightblue")
ppc_intervals(y, yrep, x = x, prob = 0.5, prob_outer = 0.9) +
  labs(x = "x covariate", y = "y", title = "Posterior predictive intervals")

# Ribbon version (polygon rather than line segments)
color_scheme_set("teal")
ppc_ribbon(y, yrep, x = x, prob = 0.5, prob_outer = 0.9)

# Grouped intervals — one facet per group
color_scheme_set("red")
ppc_intervals_grouped(y, yrep, x = x, group = group, prob = 0.5)
```

--------------------------------

### Add title to mcmc_pairs plot

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

Use the `grid_args` argument in `mcmc_pairs()` to pass options to `gridExtra::arrangeGrob()`, such as adding a title to the plot.

```R
mcmc_pairs(fit, pars = c("beta[1]", "beta[2]"), 
  grid_args = list(title = "Posterior means and correlations"))
```

--------------------------------

### Manage ggplot2 Theme for Bayespots

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Control the ggplot2 theme applied to bayesplots. Changes affect only bayesplots and not other ggplot objects. Use `bayesplot_theme_set()` to apply a theme and `bayesplot_theme_update()` to modify specific elements.

```r
library(bayesplot)
library(ggplot2)

x <- example_mcmc_draws()

# Use the built-in default (serif, 12pt)
bayesplot_theme_set()
mcmc_hist(x)

# Switch to a smaller sans-serif font
bayesplot_theme_set(theme_default(base_size = 9, base_family = "sans"))
mcmc_areas(x, regex_pars = "beta")

# Use a completely different ggplot2 theme
bayesplot_theme_set(theme_classic())
mcmc_trace(x, regex_pars = "beta")

# Update individual elements (title size, panel background)
bayesplot_theme_update(
  plot.title = element_text(size = 16, face = "bold"),
  panel.background = element_rect(fill = "gray98")
)

# Restore the default
bayesplot_theme_set()
```

--------------------------------

### Data Backend Functions

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Companion functions to major plot functions that return the underlying tidy data frame without creating a plot. Useful for custom visualizations with ggplot2.

```APIDOC
## Data Backend Functions

### `mcmc_trace_data()` / `mcmc_intervals_data()` / `mcmc_areas_data()` / `ppc_data()` / `ppc_stat_data()`

**Description:**
Every major plot function has a companion `*_data()` function that returns the underlying tidy data frame without creating a plot. These are intended for users who want to build custom visualizations with **ggplot2** on the same underlying data bayesplot would use.

### Usage:
```r
# Retrieve trace data
trace_df <- mcmc_trace_data(x, pars = c("alpha", "sigma"))
head(trace_df)

# Retrieve interval summary data
int_df <- mcmc_intervals_data(x, prob = 0.5, prob_outer = 0.9)

# Custom interval plot built from the data backend
ggplot(int_df, aes(y = parameter)) +
  geom_segment(aes(x = ll, xend = hh, yend = parameter), color = "gray60") +
  geom_segment(aes(x = l,  xend = h,  yend = parameter), linewidth = 2) +
  geom_point(aes(x = m), size = 3, shape = 21, fill = "white") +
  theme_bw() +
  labs(title = "Custom interval plot from mcmc_intervals_data()")

# Retrieve PPC data for custom plotting
ppc_df <- ppc_data(y, yrep[1:10, ])

# Retrieve test-statistic data
stat_df <- ppc_stat_data(y, yrep, stat = "median")
```
```

--------------------------------

### Control outer and inner probability intervals

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

New `prob_outer` argument and existing `prob` argument in PPC interval plots control the display of outer and inner probability intervals, consistent with `mcmc_intervals()`.

```R
ppc_intervals(y, y_rep, prob = 0.5, prob_outer = 0.9)
```

--------------------------------

### Parallel coordinates plot for MCMC draws

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

The new `mcmc_parcoord()` function creates parallel coordinates plots for MCMC draws, optionally including HMC/NUTS diagnostic information.

```R
mcmc_parcoord(fit, pars = c("beta[1]", "beta[2]"))
```

--------------------------------

### MCMC Autocorrelation Plot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Generates autocorrelation plots for MCMC draws using `mcmc_acf()`. Requires array input with chains and displays per-chain lag structure.

```R
x <- example_mcmc_draws(chains = 4)
color_scheme_set("green")
mcmc_acf(x, pars = c("alpha", "beta[1]"), lags = 20)
```

--------------------------------

### MCMC Autocorrelation Bar Plot

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Creates bar plots of autocorrelation for MCMC draws using `mcmc_acf_bar()`. Useful for visualizing lag structure across chains.

```R
mcmc_acf_bar(x, pars = c("alpha", "beta[1]"), lags = 15)
```

--------------------------------

### Deprecated argument in mcmc_trace

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

The `divergences` argument in `mcmc_trace()` is deprecated and replaced by `np` (NUTS parameters) to align with other functions.

```R
mcmc_trace(fit, pars = "beta[1]", np = nuts_params(fit, "divergent__"))
```

--------------------------------

### MCMC Density Areas

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Shades credible intervals using `mcmc_areas()`. Can display density areas and outer tails, with options for different area calculation methods.

```R
color_scheme_set("red")
mcmc_areas(
  x,
  regex_pars = "beta\[[1-3]\]",
  prob       = 0.8,
  area_method = "equal area"
) +
  labs(title = "Posterior distributions", subtitle = "with medians and 80% intervals")
```

--------------------------------

### Plot Grid

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Arrange multiple bayesplots (ggplot objects) in a shared grid, optionally aligning axes and adding titles. This function wraps `gridExtra::arrangeGrob()`.

```APIDOC
## Plot Grid

### `bayesplot_grid()`

**Description:**
Arrange multiple bayesplots (ggplot objects) in a shared grid, optionally aligning x/y axes and adding titles. Wraps `gridExtra::arrangeGrob()`.

### Usage:
```r
# Arrange plots in a grid with specified titles and layout
bayesplot_grid(
  plots     = lapply(stats, function(s) ppc_stat(y, yrep, stat = s)),
  titles    = stats,
  legends   = FALSE,
  grid_args = list(ncol = 2)
)

# Share x-axis limits across multiple interval plots in a single row
bayesplot_grid(
  mcmc_intervals(x, pars = c("alpha", "sigma")),
  mcmc_areas(x, regex_pars = "beta\[[1-3]\]"),
  xlim = c(-3, 3),
  grid_args = list(nrow = 1)
)
```
```

--------------------------------

### Histogram breaks argument

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

The `breaks` argument in histogram plots can now be used as an alternative to `binwidth` for controlling binning.

```R
mcmc_hist(log_lik, binwidth = 0.1, breaks = seq(-5, 5, by = 0.5))
```

--------------------------------

### Control bayesplot ggplot2 theme

Source: https://github.com/stan-dev/bayesplot/blob/master/NEWS.md

Use these functions to manage the ggplot theme specifically for bayesplot plots. They mirror ggplot2's theme functions but only affect bayesplot output.

```R
bayesplot_theme_set()
bayesplot_theme_get()
bayesplot_theme_update()
bayesplot_theme_replace()
```

--------------------------------

### MCMC Ridgeline Areas

Source: https://context7.com/stan-dev/bayesplot/llms.txt

Generates ridgeline plots of density areas using `mcmc_areas_ridges()`. Suitable for compact display of hierarchically related parameters.

```R
color_scheme_set("purple")
mcmc_areas_ridges(x, regex_pars = "beta", border_size = 0.75) +
  ggtitle("Hierarchical beta parameters")
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.