## ----include = FALSE----------------------------------------------------------
knitr::opts_chunk$set(
  collapse = TRUE,
  warning = FALSE,
  message = FALSE,
  comment = "#>"
)

## ----setup--------------------------------------------------------------------
library(d3po)
library(sf)
library(igraph)

## ----bar1---------------------------------------------------------------------
trade_by_continent <- d3po::trade[d3po::trade$year == 2023L, ]
trade_by_continent <- aggregate(
  trade ~ reporter_continent,
  data = d3po::trade,
  FUN = sum
)

# Assign colors to continents
my_pal <- tintin::tintin_pal()(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_bar(daes(x = reporter_continent, y = trade, color = my_pal)) %>%
  po_labels(
    x = "Continent",
    y = "Trade (USD billion)",
    title = "Total Trade by Reporter Continent in 2023"
  )

## ----bar2---------------------------------------------------------------------
trade_by_continent$color <- my_pal[trade_by_continent$reporter_continent]

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_bar(daes(x = trade, y = reporter_continent, color = color)) %>%
  po_labels(
    x = "Trade (USD billion)",
    y = "Continent",
    title = "Total Trade by Reporter Continent in 2023"
  )

## ----bar3---------------------------------------------------------------------
trade_stacked <- d3po::trade
trade_stacked <- aggregate(trade ~ reporter_continent + partner_continent, data = trade_stacked, FUN = sum)

trade_stacked$color <- ifelse(trade_stacked$partner_continent == "Africa", my_pal["Africa"], NA)
trade_stacked$color <- ifelse(trade_stacked$partner_continent == "Antarctica", my_pal["Antarctica"], trade_stacked$color)
trade_stacked$color <- ifelse(trade_stacked$partner_continent == "Asia", my_pal["Asia"], trade_stacked$color)
trade_stacked$color <- ifelse(trade_stacked$partner_continent == "Europe", my_pal["Europe"], trade_stacked$color)
trade_stacked$color <- ifelse(trade_stacked$partner_continent == "North America", my_pal["North America"], trade_stacked$color)
trade_stacked$color <- ifelse(trade_stacked$partner_continent == "Oceania", my_pal["Oceania"], trade_stacked$color)
trade_stacked$color <- ifelse(trade_stacked$partner_continent == "South America", my_pal["South America"], trade_stacked$color)

d3po(trade_stacked, width = 800, height = 600) %>%
  po_bar(daes(
    x = reporter_continent, y = trade, group = partner_continent,
    color = color, stack = TRUE
  )) %>%
  po_labels(
    x = "Reporter Continent",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter and Partner Continent in 2023"
  )

## ----bar4---------------------------------------------------------------------
d3po(trade_by_continent, width = 800, height = 600) %>%
  po_bar(daes(x = reporter_continent, y = trade, color = my_pal)) %>%
  po_labels(
    x = "Reporter Continent",
    y = "Trade (USD billion)",
    title = "Total Trade by Reporter Continent in 2023"
  ) %>%
  po_theme(axis = "#012169", tooltip = "#101418", background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE)

## ----pie1---------------------------------------------------------------------
trade_by_continent <- d3po::trade[d3po::trade$year == 2023L, ]
trade_by_continent <- aggregate(
  trade ~ reporter_continent,
  data = d3po::trade,
  FUN = sum
)

# Assign colors to continents
my_pal <- tintin::tintin_pal(option = "The Black Island")(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_pie(daes(size = trade, group = reporter_continent, color = my_pal)) %>%
  po_labels(title = "Trade Share by Reporter Continent in 2023")

## ----donut1-------------------------------------------------------------------
trade_by_continent$color <- my_pal[trade_by_continent$reporter_continent]

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_donut(daes(size = trade, group = reporter_continent, inner_radius = 0.3, color = color)) %>%
  po_labels(title = "Trade Share by Reporter Continent in 2023")

## ----pie2---------------------------------------------------------------------
d3po(trade_by_continent, width = 800, height = 600) %>%
  po_pie(daes(size = trade, group = reporter_continent, color = my_pal)) %>%
  po_labels(title = "Trade Share by Reporter Continent in 2023") %>%
  po_theme(tooltip = "#101418", background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE)

## ----area1--------------------------------------------------------------------
trade_by_continent <- d3po::trade
trade_by_continent <- aggregate(
  trade ~ year + reporter_continent,
  data = trade_by_continent,
  FUN = sum
)

# Assign colors to continents
my_pal <- tintin::tintin_pal(option = "Cigars of the Pharaoh")(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_area(daes(
    x = year, y = trade, group = reporter_continent, color = my_pal
  )) %>%
  po_labels(
    x = "Year",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter Continent in 2019 and 2023"
  )

## ----area2--------------------------------------------------------------------
trade_by_continent$color <- my_pal[trade_by_continent$reporter_continent]

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_area(daes(
    x = year, y = trade, group = reporter_continent, color = color
  )) %>%
  po_labels(
    x = "Year",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter Continent in 2019 and 2023"
  )

## ----area3--------------------------------------------------------------------
trade_by_continent$proportion <- ave(
  trade_by_continent$trade,
  trade_by_continent$year,
  FUN = function(x) x / sum(x)
)

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_area(daes(
    x = year, y = proportion, group = reporter_continent, color = my_pal, stack = TRUE
  )) %>%
  po_labels(
    x = "Year",
    y = "Proportion of Trade",
    title = "Trade Proportions by Reporter Continent in 2019 and 2023"
  )

## ----area4--------------------------------------------------------------------
d3po(trade_by_continent, width = 800, height = 600) %>%
  po_area(daes(
    x = year, y = proportion, group = reporter_continent, color = color, stack = TRUE
  )) %>%
  po_labels(
    x = "Year",
    y = "Proportion of Trade",
    title = "Trade Proportions by Reporter Continent in 2019 and 2023"
  )

## ----area5--------------------------------------------------------------------
d3po(trade_by_continent, width = 800, height = 600) %>%
  po_area(daes(
    x = year, y = trade, group = reporter_continent, color = my_pal
  )) %>%
  po_labels(
    x = "Year",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter Continent in 2019 and 2023"
  ) %>%
  po_theme(axis = "#012169", tooltip = "#101418", background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE)

## ----line1--------------------------------------------------------------------
trade_by_continent <- d3po::trade
trade_by_continent <- aggregate(
  trade ~ year + reporter_continent,
  data = trade_by_continent,
  FUN = sum
)

# Assign colors to continents
my_pal <- tintin::tintin_pal(option = "The Broken Ear")(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_line(daes(x = year, y = trade, group = reporter_continent, color = my_pal)) %>%
  po_labels(
    x = "Year",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter Continent in 2019 and 2023"
  )

## ----line2--------------------------------------------------------------------
trade_by_continent$color <- my_pal[trade_by_continent$reporter_continent]

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_line(daes(x = year, y = trade, group = reporter_continent, color = color)) %>%
  po_labels(
    x = "Year",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter Continent in 2019 and 2023"
  )

## ----line3--------------------------------------------------------------------
d3po(trade_by_continent, width = 800, height = 600) %>%
  po_line(daes(x = year, y = trade, group = reporter_continent, color = my_pal)) %>%
  po_labels(
    x = "Year",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter Continent in 2019 and 2023"
  ) %>%
  po_theme(axis = "#012169", tooltip = "#101418", background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE)

## ----scatter1-----------------------------------------------------------------
# Create a wide dataset with x = 2019 and y = 2023 trade values
trade_wide_2019 <- d3po::trade[d3po::trade$year == 2019L, c("reporter", "trade")]
trade_wide_2019 <- aggregate(trade ~ reporter, data = trade_wide_2019, FUN = sum)

trade_wide_2023 <- d3po::trade[d3po::trade$year == 2023L, c("reporter", "trade")]
trade_wide_2023 <- aggregate(trade ~ reporter, data = trade_wide_2023, FUN = sum)

trade_wide <- merge(
  trade_wide_2019,
  trade_wide_2023,
  by = "reporter",
  suffixes = c("_2019", "_2023")
)

my_pal <- tintin::tintin_pal(option = "red_rackhams_treasure")(7)

d3po(trade_wide, width = 800, height = 600) %>%
  po_scatter(daes(x = trade_2019, y = trade_2023, group = reporter, color = my_pal)) %>%
  po_labels(
    x = "Trade in 2019 (USD billion)",
    y = "Trade in 2023 (USD billion)",
    title = "Trade Volume by Country in 2019 and 2023"
  )

## ----scatter2-----------------------------------------------------------------
trade_wide$color <- sample(my_pal, nrow(trade_wide), replace = TRUE)

d3po(trade_wide, width = 800, height = 600) %>%
  po_scatter(daes(x = trade_2019, y = trade_2023, group = reporter, color = color)) %>%
  po_labels(
    x = "Trade in 2019 (USD billion)",
    y = "Trade in 2023 (USD billion)",
    title = "Trade Volume by Country in 2019 and 2023"
  )

## ----scatter3-----------------------------------------------------------------
trade_wide$size <- (trade_wide$trade_2019 + trade_wide$trade_2023) / 2

d3po(trade_wide, width = 800, height = 600) %>%
  po_scatter(daes(
    x = trade_2019, y = trade_2023,
    group = reporter, color = color, size = size
  )) %>%
  po_labels(
    x = "Trade in 2019 (USD billion)",
    y = "Trade in 2023 (USD billion)",
    title = "Trade Volume by Country in 2019 and 2023"
  )

## ----scatter4-----------------------------------------------------------------
d3po(trade_wide, width = 800, height = 600) %>%
  po_scatter(daes(x = trade_2019, y = trade_2023, group = reporter, color = my_pal)) %>%
  po_labels(
    x = "Trade in 2019 (USD billion)",
    y = "Trade in 2023 (USD billion)",
    title = "Trade Volume by Country in 2019 and 2023"
  ) %>%
  po_theme(axis = "#012169", tooltip = "#101418", background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE)

## ----box1---------------------------------------------------------------------
trade_continent <- d3po::trade
trade_continent <- aggregate(
  trade ~ reporter_continent + reporter,
  data = trade_continent,
  FUN = sum
)

my_pal <- tintin::tintin_pal(option = "Destination Moon")(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

d3po(trade_continent, width = 800, height = 600) %>%
  po_box(daes(x = reporter_continent, y = trade, color = my_pal, tooltip = reporter_continent)) %>%
  po_labels(
    x = "Continent",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter Continent"
  )

## ----box2---------------------------------------------------------------------
trade_continent$color <- my_pal[trade_continent$reporter_continent]

d3po(trade_continent, width = 800, height = 600) %>%
  po_box(daes(y = reporter_continent, x = trade, color = color, tooltip = reporter_continent)) %>%
  po_labels(
    y = "Continent",
    x = "Trade (USD billion)",
    title = "Trade Distribution by Continents with Custom Colors"
  )

## ----box3---------------------------------------------------------------------
d3po(trade_continent, width = 800, height = 600) %>%
  po_box(daes(x = reporter_continent, y = trade, color = my_pal, tooltip = reporter_continent)) %>%
  po_labels(
    x = "Continent",
    y = "Trade (USD billion)",
    title = "Trade Distribution by Reporter Continent"
  ) %>%
  po_theme(axis = "#012169", tooltip = "#101418", background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE)

## ----treemap1-----------------------------------------------------------------
trade_by_continent <- d3po::trade[d3po::trade$year == 2023L, ]
trade_by_continent <- aggregate(trade ~ reporter_continent, data = trade_by_continent, FUN = sum)

my_pal <- tintin::tintin_pal(option = "The Secret of the Unicorn")(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_treemap(daes(size = trade, group = reporter_continent, color = my_pal, tiling = "squarify")) %>%
  po_labels(title = "Trade Share by Continent in 2023")

## ----treemap2-----------------------------------------------------------------
trade_by_continent$color <- my_pal[trade_by_continent$reporter_continent]

d3po(trade_by_continent, width = 800, height = 600) %>%
  po_treemap(daes(size = trade, group = reporter_continent, color = color, tiling = "slice-dice")) %>%
  po_labels(title = "Trade Share by Continent in 2023")

## ----treemap3-----------------------------------------------------------------
trade_twolevel <- d3po::trade[d3po::trade$year == 2023L, ]
trade_twolevel <- aggregate(trade ~ reporter_continent + reporter, data = trade_twolevel, FUN = sum)
trade_twolevel$color <- my_pal[trade_twolevel$reporter_continent]

d3po(trade_twolevel, width = 800, height = 600) %>%
  po_treemap(daes(
    size = trade, group = reporter_continent, subgroup = reporter,
    color = color, tiling = "squarify"
  )) %>%
  po_labels(title = "Trade Share by Continent in 2023 (click to see the countries)")

## ----treemap4-----------------------------------------------------------------
d3po(trade_twolevel, width = 800, height = 600) %>%
  po_treemap(daes(
    size = trade, group = reporter_continent, subgroup = reporter,
    color = color, tiling = "squarify"
  )) %>%
  po_theme(background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE) %>%
  po_labels(
    align = "center-middle",
    labels = JS(
      "function(percentage, row) {
          var pct = (percentage).toFixed(2) + '%';
          // Show reporter (country) if available, otherwise show reporter_continent
          var name = (row && row.reporter) ? row.reporter : (row && row.reporter_continent ? row.reporter_continent : '');
          var count = row && (row.trade != null ? row.trade : (row.value != null ? row.value : ''));
          count = (count).toFixed(2) + 'B';
          return '<i>' + name + '</i><br/>Trade: ' + (count || '') + '<br/>Percentage: ' + pct;\n
        }"
    ),
    title = "Trade Share by Continent in 2023 (click to see the countries)",
    subtitle = JS(
      "function(_v, row) {
          // row.mode is 'aggregated' | 'flat' | 'drilled'
          if (row && row.mode === 'drilled') return 'Displaying Countries';
          return 'Displaying Continents';\
        }"
    )
  ) %>%
  po_tooltip(JS(
    "function(percentage, row) {
        var pct = (percentage).toFixed(2) + '%';

        var count = row && row.count != null ? row.count : '';
        count = (count).toFixed(2) + 'B';

        if (!row || !row.reporter) {
          var t1 = row && (row.reporter_continent || row.reporter) ? (row.reporter_continent || row.reporter) : '';
          return '<i>Continent: ' + t1 + '</i><br/>Trade: ' + count + '<br/>Percentage: ' + pct;
        }

        return '<i>Continent: ' + (row.reporter_continent || '') + '<br/>Country: ' + (row.reporter || '') +
        '</i><br/>Trade: ' + count + '<br/>Percentage: ' + pct;
      }"
  ))

## ----geomap1------------------------------------------------------------------
world <- d3po::national

# Fix geometries that cross the antimeridian (date line) to avoid horizontal lines
# This affects Russia, Fiji, and other countries spanning the 180° meridian
world$geometry <- sf::st_wrap_dateline(world$geometry, options = c("WRAPDATELINE=YES"))

total_trade <- d3po::trade[d3po::trade$year == 2023L, c("reporter", "reporter_continent", "trade")]
total_trade <- aggregate(trade ~ reporter, data = total_trade, FUN = sum)
colnames(total_trade) <- c("country", "trade")

world <- merge(
  world,
  total_trade,
  by = "country",
  all.x = TRUE,
  all.y = FALSE
)

my_pal <- tintin::tintin_pal(option = "The Calculus Affair")(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

d3po(world, width = 800, height = 600) %>%
  po_geomap(daes(group = country, size = trade, color = my_pal, tooltip = country)) %>%
  po_labels(title = "Trade Volume by Country in 2023")

## ----geomap2------------------------------------------------------------------
world$color <- my_pal[world$continent]

d3po(world, width = 800, height = 600) %>%
  po_geomap(daes(group = country, size = trade, color = color, tooltip = country)) %>%
  po_labels(title = "Trade Volume by Country in 2023")

## ----geomap3------------------------------------------------------------------
europe <- world[world$continent == "Europe", ]

# Filter to continental Europe + Iceland using bounding box
# This excludes overseas territories like Canary Islands, French Guiana, etc.
bbox <- sf::st_bbox(c(xmin = -27, ymin = 30, xmax = 40, ymax = 72), crs = sf::st_crs(europe))
europe <- sf::st_crop(europe, bbox)

europe$color <- my_pal[europe$continent]

my_color <- c("#e74c3c", "#3498db", "#2ecc71")

d3po(europe, width = 800, height = 600) %>%
  po_geomap(daes(group = country, size = trade, color = my_color, tooltip = country)) %>%
  po_labels(title = "Trade Volume by Country in 2023")

## ----geomap4------------------------------------------------------------------
d3po(europe, width = 800, height = 600) %>%
  po_geomap(daes(group = country, size = trade, gradient = TRUE, tooltip = country)) %>%
  po_labels(title = "Trade Volume by Country in 2023")

## ----geomap5------------------------------------------------------------------
d3po(europe, width = 800, height = 600) %>%
  po_geomap(daes(group = country, size = trade, color = my_color, gradient = TRUE, tooltip = country)) %>%
  po_labels(title = "Trade Volume by Country in 2023")

## ----geomap6------------------------------------------------------------------
d3po(europe, width = 800, height = 600) %>%
  po_geomap(daes(group = country, size = trade, color = my_color, gradient = TRUE, tooltip = country)) %>%
  po_labels(title = "Trade Volume by Country in 2023") %>%
  po_theme(axis = "#012169", tooltip = "#101418", background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE)

## ----network1-----------------------------------------------------------------
trade_network <- d3po::trade[d3po::trade$year == 2023L, ]
trade_network <- aggregate(trade ~ reporter_iso + partner_iso + reporter_continent + partner_continent,
  data = trade_network, FUN = sum
)

# subset to 5 largest connection per reporter country
trade_network <- do.call(
  rbind,
  lapply(
    split(trade_network, trade_network$reporter_iso),
    function(df) head(df[order(-df$trade), ], 5)
  )
)

# Create vertex (node) attributes for coloring and sizing
# Get unique countries with their continents and trade volumes
vertices <- unique(rbind(
  data.frame(
    name = trade_network$reporter_iso,
    continent = trade_network$reporter_continent,
    stringsAsFactors = FALSE
  ),
  data.frame(
    name = trade_network$partner_iso,
    continent = trade_network$partner_continent,
    stringsAsFactors = FALSE
  )
))

# Remove duplicates
vertices <- vertices[!duplicated(vertices$name), ]

# Calculate total trade volume per country (as reporter)
trade_volume <- aggregate(trade ~ reporter_iso, data = trade_network, FUN = sum)
colnames(trade_volume) <- c("name", "trade_volume")

# Merge trade volume with vertices
vertices <- merge(vertices, trade_volume, by = "name", all.x = TRUE)
vertices$trade_volume[is.na(vertices$trade_volume)] <- 0

# Assign colors to continents
my_pal <- tintin::tintin_pal(option = "The Blue Lotus")(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

# Add color column based on continent
vertices$color <- my_pal[vertices$continent]

# Create igraph object with vertex attributes
g <- graph_from_data_frame(trade_network, directed = TRUE, vertices = vertices)

# Create the network visualization
d3po(g, width = 800, height = 600) %>%
  po_network(daes(size = trade_volume, color = color, layout = "fr")) %>%
  po_labels(title = "Trade Network by Country in 2023")

## ----network2-----------------------------------------------------------------
# Use a different color palette
my_pal <- tintin::tintin_pal(option = "Explorers on the Moon")(7)

names(my_pal) <- c(
  "Africa", "Antarctica", "Asia",
  "Europe", "North America", "Oceania", "South America"
)

# Update colors with new palette
vertices$color <- my_pal[vertices$continent]

# Create network with Kamada-Kawai layout
d3po(g, width = 800, height = 600) %>%
  po_network(daes(size = trade_volume, color = color, layout = "kk")) %>%
  po_labels(title = "Trade Network by Country in 2023")

## ----network3-----------------------------------------------------------------
d3po(g, width = 800, height = 600) %>%
  po_network(daes(size = trade_volume, color = color, layout = "fr")) %>%
  po_labels(title = "Trade Network by Country in 2023") %>%
  po_theme(tooltip = "#101418", background = "#cccccc") %>%
  po_font("Liberation Serif", 12, "uppercase") %>%
  po_download(FALSE)