R/extra.R

In seananderson/gfplot: Data Extraction and Plotting for DFO PBS Groundfish Stocks

plot_cpue_predictions <- function(dat, model_version = "Combined", scale = FALSE) {
  if (scale) {
    dat <- dat %>%
      group_by(.data$formula_version, .data$model, .data$area) %>%
      mutate(geo_mean = exp(mean(log(.data$est)))) %>%
      mutate(upr = .data$upr / .data$geo_mean, lwr = .data$lwr / .data$geo_mean,
        est = .data$est / .data$geo_mean) %>%
      ungroup()
  }

  unstandardized <- dat %>%
    filter(.data$formula_version == "Unstandardized") %>%
    rename(est_unstandardized = .data$est,
      lwr_unstandardized = .data$lwr, upr_unstandardized = .data$upr) %>%
    select(.data$year, .data$area, .data$model,
      .data$est_unstandardized, .data$lwr_unstandardized,
      .data$upr_unstandardized)

  temp <- dat %>%
    filter(.data$formula_version != "Unstandardized") %>%
    left_join(unstandardized, by = c("area", "year", "model")) %>%
    filter(model == model_version) %>%
    mutate(formula_version = gsub("\\+ ", " ", .data$formula_version)) %>%
    mutate(
      formula_version =
        gsub("Full standardization", "All variables", .data$formula_version)
    ) %>%
    mutate(
      formula_version =
        forcats::fct_relevel(.data$formula_version, "All variables", after = Inf)
    )
  temp %>%
    ggplot(aes_string("year", "est", ymin = "upr", ymax = "lwr")) +
    geom_line(aes_string(y = "est_unstandardized"), colour = "grey30", lty = 1) +
    ggplot2::geom_ribbon(
      aes_string(ymin = "lwr_unstandardized", ymax = "upr_unstandardized"),
      fill = "#00000030"
    ) +
    ggplot2::geom_ribbon(alpha = 0.4, fill = "red") +
    geom_line(colour = "red") +
    theme_pbs() +
    facet_grid(area ~ formula_version, scales = "free_y") +
    labs(
      y = if (!scale) "CPUE (kg/hour)" else "CPUE (kg/hour) divided\nby geometric mean",
      x = ""
    ) +
    ylim(0, NA) +
    guides(colour = "none", fill = "none")
}