Electricity plot

electricity_generation_plot.R

#' @examples
#' renv::install("isocountry")
#' renv::install("readr")
#' renv::install("dplyr")
#' renv::install("forcats")
#' renv::install("rlang")
#' renv::install("glue")
#' renv::install("readr")
#' renv::install("clauswilke/ggtextures")

library(isocountry)
library(ggplot2)
library(ggtextures)
library(dplyr)
library(forcats)
library(glue)
library(rlang)
library(readr)

country_emoji <- function(country_codes) {
  vapply(
    country_codes,
    function(country_code) {
      if (is.null(country_code) || isTRUE(is.na(country_code))) {
        return("")
      }
      intToUtf8(127397 + strtoi(charToRaw(toupper(country_code)), 16L))
    },
    character(1)
  )
}

#
# Order of Fuel Types (from most pollutent to least)
levels_factor <- c(
  "Coal", "Gas", "Other Fossil", "Nuclear", "Bioenergy", "Hydro", "Wind",
  "Solar", "Other Renewables"
)

levels_ordered <- set_names(seq_along(levels_factor), levels_factor)

colors_fuel <- c(
  "#060607", "#292f34" , "#eeeeee", # Fossil
  "#56B4E9", # Nuclear
  "#E69F00", "#CC79A7", "#F0E442", "#D55E00", "#0072B2" # Renewables
)

textured_mask <- file.path(
  "https://gist.githubusercontent.com",
  "averissimo/8804d3ffed0613f9f912c4ef3465fbc1/raw",
  "50b92ab34c8d8c7e73a4d5cfeb847b35b75ff417/noisy-texture.png"
)

text_color <- "#36454f"

#
#
url <- file.path(
  "https://ember-climate.org",
  "app/uploads/2022/07/yearly_full_release_long_format.csv"
)

dat <- read_csv(url) |>
  filter(.data[["Area type"]] == "Country") |>
  filter(.data[["EU"]] == 1) |>
  left_join(
    isocountry |>
      select(iso_2 = "alpha_2", "Country code" = "alpha_3") |>
      bind_rows(
        tibble(
          iso_2 = "XK",
          "Country code" = "XKX"
        )
      )
  ) |>
  relocate(iso_2, .after = "Country code") |>
  # filter(.data[["Country code"]] %in% c("DEU", "PRT")) |>
  filter(.data$Year == "2023")

# Filtered data to work with
dat_filtered <- dat |>
  filter(.data$Unit != "%") |>
  filter(
    .data$Category == "Electricity generation",
    .data$Subcategory == "Fuel"
  ) |>
  mutate(total = sum(.data$Value)) |>
  group_by(.data$Area) |>
  mutate(
    subtotal = sum(.data$Value),
    polluting_order = sum(if_else(.data$Variable %in% c("Coal", "Gas", "Other Fossil"), .data$Value, 0))
  ) |>
  ungroup()  |>
  mutate(
    polluting_order = .data$polluting_order / .data$subtotal,
    relative = .data$subtotal / .data$total,
    Variable = fct(.data$Variable, levels = levels_factor)
  ) |>
  select(-c("Area type", "Ember region", "EU", "OECD", "G7", "ASEAN"))

# Data for first plot
dat_plot <- dat_filtered |>
  mutate(
    Value = .data$Value / .data$subtotal * 100,
    emoji = unlist(country_emoji(.data$iso_2)),
    .after = "iso_2"
  ) |>
  mutate(
    relative_rounded = relative * 100,
    relative_rounded = if_else(relative_rounded >= 1, round(relative_rounded, digits = 1), round(relative_rounded, digits = 2)),
    label = glue("{.data$Area} {.data$emoji}"),
    .before = "Area"
  ) |>
  as_tibble() |>
  mutate(
    label = fct_reorder(.data$label, .data$polluting_order)
  )

# ggplot(dat_plot) +
#   aes(x = label, y = Value, fill = Variable) +
#   geom_bar(stat = "identity", position = "stack") +
#   coord_flip() +
#   scale_fill_manual(values = colors_fuel) +
#   theme_minimal()

# Data for manual rectangular plot
dat_plot_2 <- dat_plot |>
  select(label, Area, Variable, Value, subtotal, total, relative, polluting_order, emoji) |>
  mutate(levels_order = levels_ordered[.data$Variable]) |>
  group_by(Area) |>
  arrange(desc(polluting_order), subtotal, Area, levels_order) |>
  mutate(
    ymax = cumsum(lag(.data$Value, default = 0)), ymax = cumsum(.data$Value),
    ymin = ymax - .data$Value
  ) |>
  rowwise() |>
  mutate(
    # subtotal_norm = log(1 + subtotal)
    subtotal_norm = max(subtotal, 50) # Minimal cap for the bar height
  ) |>
  group_by(Area) |>
  mutate(x = cumsum(if_else(row_number() == n(), subtotal_norm, 0))) |>
  ungroup() |>
  mutate(
    xmin = cumsum(lag(x, default = 0)),
    xmax = xmin + subtotal_norm
  )

# Flip the plot manually
dat_plot_2_min <- dat_plot_2 |>
  group_by(Area, label, relative, subtotal, emoji) |>
  summarise(
    ymin = min(ymin),
    ymax = max(ymax),
    xmin = first(xmin),
    xmax = last(xmax),
    .groups = "keep"
  )

dat_plot_2_ready <- dat_plot_2 |>
  rename(
    xmin = ymin,
    xmax = ymax,
    ymin = xmin,
    ymax = xmax
  ) |>
  select(label, Variable, xmin, xmax, ymin, ymax, subtotal, emoji) |>
  filter(xmin != xmax) # Bug in geom_textured_*

dat_plot_2_min_ready <- dat_plot_2_min |> rename(
  xmin = ymin,
  xmax = ymax,
  ymin = xmin,
  ymax = xmax
) |>
  group_by(Area, label, relative) |>
  mutate(
    y = ymin + (ymax - ymin) / 2,
    label_2 = glue("{emoji} {.data$subtotal} TWh ({round(relative * 100, digits = 2)}%)")
  )

ggplot(dat_plot_2_ready) +
  aes(xmin = xmin, xmax = xmax, ymin = ymin, ymax = ymax, fill = Variable) +
  geom_vline(xintercept = seq(0, 100, 25), color = "#777777ff", linetype = "dashed") +
  geom_rect() +
  geom_textured_rect(image = textured_mask, img_width = unit(.5, "in"), color = "#ffffff") +
  geom_rect(data = dat_plot_2_min_ready, fill = "transparent", color = "white") +
  geom_vline(xintercept = seq(0, 100, 25), color = "#eeeeee88", linetype = "dashed") +
  scale_x_continuous(breaks = seq(0, 100, 25)) +
  geom_text(
    data = dat_plot_2_min_ready,
    aes(y = y, label = label),
    x = -.5,
    size = 4,
    inherit.aes = FALSE,
    hjust = 1,
    nudge_x = 2,
    xlim = c(-15, 0),
    force_pull = 10,
    max.overlaps = Inf,
    min.segment.length = 0,
    segment.alpha = .4,
    color = text_color
  ) +
  geom_text(
    data = dat_plot_2_min_ready,
    aes(y = y, label = label_2),
    x = 100.5,
    size = 4,
    inherit.aes = FALSE,
    hjust = 0,
    color = text_color
  ) +
  geom_text(
    data = data.frame(label = glue("Country total in TWh (% of EU)")),
    aes(label = label),
    y = max(dat_plot_2_ready$ymax) + 10,
    x = 100.5,
    hjust = 0,
    size = 3,
    inherit.aes = FALSE,
    color = text_color
  ) +
  scale_fill_manual(values = colors_fuel) +
  theme_minimal() +
  theme(
    axis.title.y = element_blank(),
    axis.text.y = element_blank(),
    axis.ticks.y = element_blank(),
    legend.position = "top",
    panel.grid.major = element_blank(),
    panel.grid.minor = element_blank(),
    text = element_text(family="Helvetica-Narrow"),
    plot.margin = margin(0, 0, 0, 0, "pt"),
    legend.title = element_text(
      family = "Helvetica-Narrow",
      color = text_color,
      size = 10
    ),
    legend.text = element_text(
      family = "Helvetica-Narrow",
      color = text_color,
      size = 10
    ),
    title = element_text(family = "Helvetica-Narrow", color = text_color)
  ) +
  expand_limits(x = c(-6, 113)) +
  # change legend so it only has 1 row
  guides(fill = guide_legend(nrow = 1)) +
  theme(plot.margin = margin(1, 1, 1, 1, "cm")) +
  labs(
    title = glue("2023 {country_emoji('eu')} Rankings: Electricity generation by fuel type in EU countries"),
    subtitle = "Height of bars represents electricity generation per country in 2023",
    fill = "Fuel type:",
    x = "Share of electricity generation (%)",
    caption = "Source: Anual electricity data, Ember."
  )

noisy-texture.png

plot.png