Visualizing vitalBayes Results

Overview

vitalBayes provides publication-ready visualization functions for all model types, with built-in support for colorblind-safe palettes and multilingual labels.

Color Palettes

library(vitalBayes)

# List available palettes
list_vital_palettes()

# Default retro-wave palette
vital_colors(5, "default")

# Two-sex palette (pink/cyan)
vital_colors(2, "sex")

# Colorblind-safe alternatives
vital_colors(5, "okabe_ito")
vital_colors(2, "sex_cb")        # Orange/blue for sex
vital_colors(5, "viridis")
vital_colors(5, "tol_bright")

# Check if a palette is colorblind-safe
is_colorblind_safe("okabe_ito")  # TRUE
is_colorblind_safe("sex")        # FALSE

The vitalBayes Theme

All plots use theme_vital(), a clean ggplot2 theme optimized for publications:

library(ggplot2)

# Use standalone
ggplot(mtcars, aes(mpg, wt)) +
 geom_point() +
 theme_vital()

# Customize
ggplot(mtcars, aes(mpg, wt)) +
 geom_point() +
 theme_vital(base_size = 14) +
 theme(legend.position = "bottom")

plot_birth_ogive()

Visualize birth size transition probability:

library(data.table)

# Load simulated data
data(growth_data)

# Fit birth model
birth_fit <- fit_bayesian_birth(
 embryo_lts = growth_data[embryo == TRUE, fl],
 free_swimming_lts = growth_data[embryo == FALSE, fl]
)

# Basic birth ogive plot
plot_birth_ogive(
 fit = birth_fit,
 embryo_lengths = growth_data[embryo == TRUE, fl],
 freeswim_lengths = growth_data[embryo == FALSE, fl]
)

# Customized
plot_birth_ogive(
 fit = birth_fit,
 embryo_lengths = growth_data[embryo == TRUE, fl],
 freeswim_lengths = growth_data[embryo == FALSE, fl],
 show_data = TRUE,
 show_b50_line = TRUE,
 colors = vital_colors(1, "sunset"),
 x_lab = "Fork Length (cm)",
 title = "Birth Size Transition"
)

plot_maturity_ogive()

Visualize maturity probability curves:

# Prepare maturity data
mat_data <- growth_data[embryo == FALSE & !is.na(mat)]

# Fit two-sex maturity model
L50_fit <- fit_bayesian_maturity(
 maturity = "mat",
 lt = "fl",
 sex = "sex",
 data = mat_data,
 use_pooling = TRUE
)

# Basic maturity ogive plot
plot_maturity_ogive(
 fit = L50_fit,
 type = "length",
 data = mat_data,
 x_col = "fl",
 maturity_col = "mat",
 sex_col = "sex"
)

Maturity Ogive Arguments

plot_maturity_ogive(
 fit = L50_fit,
 type = "length",           # or "age"
 data = mat_data,
 
 # Data columns
 x_col = "fl",
 maturity_col = "mat",
 sex_col = "sex",
 
 # Display options
 show_data = TRUE,          # Overlay observed points
 show_rug = TRUE,           # Rug plot at top/bottom
 show_x50_line = TRUE,      # Vertical line at x50
 
 # Appearance
 colors = NULL,             # Auto from vital_colors("sex")
 colorblind = FALSE,        # Use colorblind-safe palette
 facet_sex = TRUE,          # Separate panels by sex
 
 # Labels
 x_lab = "Fork Length (cm)",
 y_lab = "Probability of Maturity",
 title = "Length-at-Maturity Ogive"
)

Multilingual Labels

# French
plot_maturity_ogive(
 fit = L50_fit,
 type = "length",
 data = mat_data,
 sex_labels = c("female" = "Femelle", "male" = "Mâle"),
 x_lab = "Longueur à la fourche (cm)",
 y_lab = "Probabilité de maturité"
)

# Spanish
plot_maturity_ogive(
 fit = L50_fit,
 type = "length",
 data = mat_data,
 sex_labels = c("female" = "Hembra", "male" = "Macho"),
 x_lab = "Longitud furcal (cm)",
 y_lab = "Probabilidad de madurez"
)

# Japanese
plot_maturity_ogive(
 fit = L50_fit,
 type = "length",
 data = mat_data,
 sex_labels = c("female" = "メス", "male" = "オス"),
 x_lab = "尾叉長 (cm)",
 y_lab = "成熟確率"
)

plot_growth_curve()

Visualize fitted growth curves with uncertainty:

# Prepare growth data
gdata <- growth_data[embryo == FALSE & !is.na(age)]

# Fit growth model
growth_fit <- fit_bayesian_growth(
 lt = "fl",
 age = "age",
 sex = "sex",
 data = gdata,
 model = "vb",
 k_based = FALSE,
 L50_fit = L50_fit
)

# Basic growth curve plot
plot_growth_curve(
 fit = growth_fit,
 data = gdata,
 age_col = "age",
 length_col = "fl",
 sex_col = "sex"
)

Growth Curve Arguments

plot_growth_curve(
 fit = growth_fit,
 data = gdata,
 
 # Data columns
 age_col = "age",
 length_col = "fl",
 sex_col = "sex",
 
 # Prediction range
 age_range = c(0, 50),      # Override auto-detection
 n_points = 100,            # Points for smooth curve
 
 # Display options
 show_data = TRUE,          # Overlay observed points
 show_50_ci = TRUE,         # Show 50% credible interval
 facet_sex = TRUE,          # Separate panels by sex
 
 # Appearance
 data_alpha = 0.4,
 ribbon_alpha = 0.2,
 ribbon_alpha_50 = 0.3,
 line_width = 1.2,
 colors = NULL,             # Auto from vital_colors("sex")
 colorblind = FALSE,
 
 # Model specification (must match fitting call)
 k_based = FALSE,
 which_model = 1,           # 1=VB, 2=Gompertz, 3=Logistic
 
 # Labels
 x_lab = "Age (years)",
 y_lab = "Fork Length (cm)",
 title = NULL
)

Additional Layers

Add custom ggplot2 layers:

library(ggplot2)

plot_growth_curve(
 fit = growth_fit,
 data = gdata,
 additional_layers = list(
   # Add horizontal line at L50
   geom_hline(yintercept = 75, linetype = "dashed", color = "gray50"),
   # Add annotation
   annotate("text", x = 5, y = 77, label = "L50", size = 3)
 ),
 theme_args = list(
   legend.position = "bottom",
   plot.title = element_text(hjust = 0.5)
 )
)

compare_growth_models()

Side-by-side comparison of multiple growth models:

# Fit multiple models
mat_aged <- growth_data[embryo == FALSE & !is.na(mat) & !is.na(age)]

t50_fit <- fit_bayesian_maturity(
 maturity = "mat", age = "age", sex = "sex",
 data = mat_aged
)

vb_fit <- fit_bayesian_growth(
 lt = "fl", age = "age", sex = "sex", data = gdata,
 model = "vb", k_based = FALSE, L50_fit = L50_fit, t50_fit = t50_fit
)

gomp_fit <- fit_bayesian_growth(
 lt = "fl", age = "age", sex = "sex", data = gdata,
 model = "gompertz", k_based = FALSE, L50_fit = L50_fit, t50_fit = t50_fit
)

logis_fit <- fit_bayesian_growth(
 lt = "fl", age = "age", sex = "sex", data = gdata,
 model = "logistic", k_based = FALSE, L50_fit = L50_fit, t50_fit = t50_fit
)

# Compare visually
compare_growth_models(
 "von Bertalanffy" = vb_fit,
 "Gompertz" = gomp_fit,
 "Logistic" = logis_fit,
 data = gdata,
 age_col = "age",
 length_col = "fl",
 sex_col = "sex",
 k_based_list = c(FALSE, FALSE, FALSE),
 which_model_list = c(1, 2, 3)
)

# Colorblind-safe
compare_growth_models(
 "von Bertalanffy" = vb_fit,
 "Gompertz" = gomp_fit,
 data = gdata,
 colorblind = TRUE
)

plot_residuals()

Diagnostic residual plots for growth models:

# All diagnostic plots
plot_residuals(
 fit = growth_fit,
 data = gdata,
 age_col = "age",
 length_col = "fl",
 sex_col = "sex",
 type = "all"              # Combined panel
)

# Individual plot types
plot_residuals(fit = growth_fit, data = gdata, type = "fitted")
plot_residuals(fit = growth_fit, data = gdata, type = "qq")
plot_residuals(fit = growth_fit, data = gdata, type = "histogram")
plot_residuals(fit = growth_fit, data = gdata, type = "age")

Saving Publication Figures

library(patchwork)

# Create multi-panel figure
p1 <- plot_maturity_ogive(L50_fit, type = "length", data = mat_data)
p2 <- plot_growth_curve(growth_fit, data = gdata)

combined <- p1 / p2 + plot_annotation(tag_levels = "A")

# Save at publication resolution
ggsave("Figure1.pdf", combined, width = 10, height = 12)
ggsave("Figure1.png", combined, width = 10, height = 12, dpi = 300)

Using Different Datasets for Demonstration

The package includes multiple datasets for different visualization needs:

# Balanced data (growth_data) - best for general demonstrations
data(growth_data)

# Imbalanced data - shows pooling effects on uncertainty bands
data(imbalanced_data)
mat_imbal <- imbalanced_data[embryo == FALSE & !is.na(mat)]

L50_imbal <- fit_bayesian_maturity(
 maturity = "mat", lt = "fl", sex = "sex",
 data = mat_imbal, use_pooling = TRUE
)

# Note the wider bands for males (sparse sex)
plot_maturity_ogive(
 fit = L50_imbal,
 type = "length",
 data = mat_imbal,
 title = "Maturity Ogive with Imbalanced Sex Ratio (150F, 34M)"
)